1CHECK_POSTGRES(1)     User Contributed Perl Documentation    CHECK_POSTGRES(1)
2
3
4

NAME

6       check_postgres.pl - a Postgres monitoring script for Nagios, MRTG,
7       Cacti, and others
8
9       This documents describes check_postgres.pl version 2.24.0
10

SYNOPSIS

12         ## Create all symlinks
13         check_postgres.pl --symlinks
14
15         ## Check connection to Postgres database 'pluto':
16         check_postgres.pl --action=connection --db=pluto
17
18         ## Same things, but using the symlink
19         check_postgres_connection --db=pluto
20
21         ## Warn if > 100 locks, critical if > 200, or > 20 exclusive
22         check_postgres_locks --warning=100 --critical="total=200:exclusive=20"
23
24         ## Show the current number of idle connections on port 6543:
25         check_postgres_txn_idle --port=6543 --output=simple
26
27         ## There are many other actions and options, please keep reading.
28
29         The latest news and documentation can always be found at:
30         https://bucardo.org/check_postgres/
31

DESCRIPTION

33       check_postgres.pl is a Perl script that runs many different tests
34       against one or more Postgres databases. It uses the psql program to
35       gather the information, and outputs the results in one of three
36       formats: Nagios, MRTG, or simple.
37
38   Output Modes
39       The output can be changed by use of the "--output" option. The default
40       output is nagios, although this can be changed at the top of the script
41       if you wish. The current option choices are nagios, mrtg, and simple.
42       To avoid having to enter the output argument each time, the type of
43       output is automatically set if no --output argument is given, and if
44       the current directory has one of the output options in its name. For
45       example, creating a directory named mrtg and populating it with
46       symlinks via the --symlinks argument would ensure that any actions run
47       from that directory will always default to an output of "mrtg" As a
48       shortcut for --output=simple, you can enter --simple, which also
49       overrides the directory naming trick.
50
51       Nagios output
52
53       The default output format is for Nagios, which is a single line of
54       information, along with four specific exit codes:
55
56       0 (OK)
57       1 (WARNING)
58       2 (CRITICAL)
59       3 (UNKNOWN)
60
61       The output line is one of the words above, a colon, and then a short
62       description of what was measured. Additional statistics information, as
63       well as the total time the command took, can be output as well: see the
64       documentation on the arguments --showperf, --perflimit, and --showtime.
65
66       MRTG output
67
68       The MRTG output is four lines, with the first line always giving a
69       single number of importance.  When possible, this number represents an
70       actual value such as a number of bytes, but it may also be a 1 or a 0
71       for actions that only return "true" or "false", such as
72       check_postgres_version.  The second line is an additional stat and is
73       only used for some actions. The third line indicates an "uptime" and is
74       not used. The fourth line is a description and usually indicates the
75       name of the database the stat from the first line was pulled from, but
76       may be different depending on the action.
77
78       Some actions accept an optional --mrtg argument to further control the
79       output.
80
81       See the documentation on each action for details on the exact MRTG
82       output for each one.
83
84       Simple output
85
86       The simple output is simply a truncated version of the MRTG one, and
87       simply returns the first number and nothing else. This is very useful
88       when you just want to check the state of something, regardless of any
89       threshold. You can transform the numeric output by appending KB, MB,
90       GB, TB, or EB to the output argument, for example:
91
92         --output=simple,MB
93
94       Cacti output
95
96       The Cacti output consists of one or more items on the same line, with a
97       simple name, a colon, and then a number. At the moment, the only action
98       with explicit Cacti output is 'dbstats', and using the --output option
99       is not needed in this case, as Cacti is the only output for this
100       action. For many other actions, using --simple is enough to make Cacti
101       happy.
102

DATABASE CONNECTION OPTIONS

104       All actions accept a common set of database options.
105
106       -H NAME or --host=NAME
107           Connect to the host indicated by NAME. Can be a comma-separated
108           list of names. Multiple host arguments are allowed. If no host is
109           given, defaults to the "PGHOST" environment variable or no host at
110           all (which indicates using a local Unix socket). You may also use
111           "--dbhost".
112
113       -p PORT or --port=PORT
114           Connects using the specified PORT number. Can be a comma-separated
115           list of port numbers, and multiple port arguments are allowed. If
116           no port number is given, defaults to the "PGPORT" environment
117           variable. If that is not set, it defaults to 5432. You may also use
118           "--dbport"
119
120       -db NAME or --dbname=NAME
121           Specifies which database to connect to. Can be a comma-separated
122           list of names, and multiple dbname arguments are allowed. If no
123           dbname option is provided, defaults to the "PGDATABASE" environment
124           variable.  If that is not set, it defaults to 'postgres' if psql is
125           version 8 or greater, and 'template1' otherwise.
126
127       -u USERNAME or --dbuser=USERNAME
128           The name of the database user to connect as. Can be a comma-
129           separated list of usernames, and multiple dbuser arguments are
130           allowed. If this is not provided, it defaults to the "PGUSER"
131           environment variable, otherwise it defaults to 'postgres'.
132
133       --dbpass=PASSWORD
134           Provides the password to connect to the database with. Use of this
135           option is highly discouraged.  Instead, one should use a .pgpass or
136           pg_service.conf file.
137
138       --dbservice=NAME
139           The name of a service inside of the pg_service.conf file. Before
140           version 9.0 of Postgres, this is a global file, usually found in
141           /etc/pg_service.conf. If you are using version 9.0 or higher of
142           Postgres, you can use the file ".pg_service.conf" in the home
143           directory of the user running the script, e.g. nagios.
144
145           This file contains a simple list of connection options. You can
146           also pass additional information when using this option such as
147           --dbservice="maindatabase sslmode=require"
148
149           The documentation for this file can be found at
150           https://www.postgresql.org/docs/current/static/libpq-pgservice.html
151
152       The database connection options can be grouped: --host=a,b --host=c
153       --port=1234 --port=3344 would connect to a-1234, b-1234, and c-3344.
154       Note that once set, an option carries over until it is changed again.
155
156       Examples:
157
158         --host=a,b --port=5433 --db=c
159         Connects twice to port 5433, using database c, to hosts a and b: a-5433-c b-5433-c
160
161         --host=a,b --port=5433 --db=c,d
162         Connects four times: a-5433-c a-5433-d b-5433-c b-5433-d
163
164         --host=a,b --host=foo --port=1234 --port=5433 --db=e,f
165         Connects six times: a-1234-e a-1234-f b-1234-e b-1234-f foo-5433-e foo-5433-f
166
167         --host=a,b --host=x --port=5432,5433 --dbuser=alice --dbuser=bob -db=baz
168         Connects three times: a-5432-alice-baz b-5433-alice-baz x-5433-bob-baz
169
170         --dbservice="foo" --port=5433
171         Connects using the named service 'foo' in the pg_service.conf file, but overrides the port
172

OTHER OPTIONS

174       Other options include:
175
176       --action=NAME
177           States what action we are running. Required unless using a
178           symlinked file, in which case the name of the file is used to
179           figure out the action.
180
181       --warning=VAL or -w VAL
182           Sets the threshold at which a warning alert is fired. The valid
183           options for this option depends on the action used.
184
185       --critical=VAL or -c VAL
186           Sets the threshold at which a critical alert is fired. The valid
187           options for this option depends on the action used.
188
189       -t VAL or --timeout=VAL
190           Sets the timeout in seconds after which the script will abort
191           whatever it is doing and return an UNKNOWN status. The timeout is
192           per Postgres cluster, not for the entire script. The default value
193           is 10; the units are always in seconds.
194
195       --assume-standby-mode
196           If specified, first the check if server in standby mode will be
197           performed (--datadir is required), if so, all checks that require
198           SQL queries will be ignored and "Server in standby mode" with OK
199           status will be returned instead.
200
201           Example:
202
203               postgres@db$./check_postgres.pl --action=version --warning=8.1 --datadir /var/lib/postgresql/8.3/main/ --assume-standby-mode
204               POSTGRES_VERSION OK:  Server in standby mode | time=0.00
205
206       --assume-prod
207           If specified, check if server in production mode is performed
208           (--datadir is required).  The option is only relevant for
209           ("symlink: check_postgres_checkpoint").
210
211           Example:
212
213               postgres@db$./check_postgres.pl --action=checkpoint --datadir /var/lib/postgresql/8.3/main/ --assume-prod
214               POSTGRES_CHECKPOINT OK: Last checkpoint was 72 seconds ago | age=72;;300 mode=MASTER
215
216       --assume-async
217           If specified, indicates that any replication between servers is
218           asynchronous.  The option is only relevant for ("symlink:
219           check_postgres_same_schema").
220
221           Example:
222               postgres@db$./check_postgres.pl --action=same_schema
223           --assume-async --dbhost=star,line
224
225       -h or --help
226           Displays a help screen with a summary of all actions and options.
227
228       --man
229           Displays the entire manual.
230
231       -V or --version
232           Shows the current version.
233
234       -v or --verbose
235           Set the verbosity level. Can call more than once to boost the
236           level. Setting it to three or higher (in other words, issuing "-v
237           -v -v") turns on debugging information for this program which is
238           sent to stderr.
239
240       --showperf=VAL
241           Determines if we output additional performance data in standard
242           Nagios format (at end of string, after a pipe symbol, using
243           name=value).  VAL should be 0 or 1. The default is 1. Only takes
244           effect if using Nagios output mode.
245
246       --perflimit=i
247           Sets a limit as to how many items of interest are reported back
248           when using the showperf option. This only has an effect for actions
249           that return a large number of items, such as table_size. The
250           default is 0, or no limit. Be careful when using this with the
251           --include or --exclude options, as those restrictions are done
252           after the query has been run, and thus your limit may not include
253           the items you want. Only takes effect if using Nagios output mode.
254
255       --showtime=VAL
256           Determines if the time taken to run each query is shown in the
257           output. VAL should be 0 or 1. The default is 1. No effect unless
258           showperf is on.  Only takes effect if using Nagios output mode.
259
260       --test
261           Enables test mode. See the "TEST MODE" section below.
262
263       --PGBINDIR=PATH
264           Tells the script where to find the psql binaries. Useful if you
265           have more than one version of the PostgreSQL executables on your
266           system, or if there are not in your path. Note that this option is
267           in all uppercase. By default, this option is not allowed. To enable
268           it, you must change the $NO_PSQL_OPTION near the top of the script
269           to 0. Avoid using this option if you can, and instead use
270           environment variable c<PGBINDIR> or hard-coded $PGBINDIR variable,
271           also near the top of the script, to set the path to the PostgreSQL
272           to use.
273
274       --PSQL=PATH
275           (deprecated, this option may be removed in a future release!)
276           Tells the script where to find the psql program. Useful if you have
277           more than one version of the psql executable on your system, or if
278           there is no psql program in your path. Note that this option is in
279           all uppercase. By default, this option is not allowed. To enable
280           it, you must change the $NO_PSQL_OPTION near the top of the script
281           to 0. Avoid using this option if you can, and instead hard-code
282           your psql location into the $PSQL variable, also near the top of
283           the script.
284
285       --symlinks
286           Creates symlinks to the main program for each action.
287
288       --output=VAL
289           Determines the format of the output, for use in various programs.
290           The default is 'nagios'. Available options are 'nagios', 'mrtg',
291           'simple' and 'cacti'.
292
293       --mrtg=VAL
294           Used only for the MRTG or simple output, for a few specific
295           actions.
296
297       --debugoutput=VAL
298           Outputs the exact string returned by psql, for use in debugging.
299           The value is one or more letters, which determine if the output is
300           displayed or not, where 'a' = all, 'c' = critical, 'w' = warning,
301           'o' = ok, and 'u' = unknown. Letters can be combined.
302
303       --get_method=VAL
304           Allows specification of the method used to fetch information for
305           the "new_version_cp", "new_version_pg", "new_version_bc",
306           "new_version_box", and "new_version_tnm" checks.  The following
307           programs are tried, in order, to grab the information from the web:
308           GET, wget, fetch, curl, lynx, links. To force the use of just one
309           (and thus remove the overhead of trying all the others until one of
310           those works), enter one of the names as the argument to get_method.
311           For example, a BSD box might enter the following line in their
312           ".check_postgresrc" file:
313
314             get_method=fetch
315
316       --language=VAL
317           Set the language to use for all output messages. Normally, this is
318           detected by examining the environment variables LC_ALL,
319           LC_MESSAGES, and LANG, but setting this option will override any
320           such detection.
321

ACTIONS

323       The action to be run is selected using the --action flag, or by using a
324       symlink to the main file that contains the name of the action inside of
325       it. For example, to run the action "timesync", you may either issue:
326
327         check_postgres.pl --action=timesync
328
329       or use a program named:
330
331         check_postgres_timesync
332
333       All the symlinks are created for you in the current directory if use
334       the option --symlinks:
335
336         perl check_postgres.pl --symlinks
337
338       If the file name already exists, it will not be overwritten. If the
339       file exists and is a symlink, you can force it to overwrite by using
340       "--action=build_symlinks_force".
341
342       Most actions take a --warning and a --critical option, indicating at
343       what point we change from OK to WARNING, and what point we go to
344       CRITICAL. Note that because criticals are always checked first, setting
345       the warning equal to the critical is an effective way to turn warnings
346       off and always give a critical.
347
348       The current supported actions are:
349
350   archive_ready
351       ("symlink: check_postgres_archive_ready") Checks how many WAL files
352       with extension .ready exist in the pg_xlog/archive_status directory
353       (PostgreSQL 10 and later: pg_wal/archive_status), which is found off of
354       your data_directory. If the --lsfunc option is not used then this
355       action must be run as a superuser, in order to access the contents of
356       the pg_xlog/archive_status directory. The minimum version to use this
357       action is Postgres 8.1. The --warning and --critical options are simply
358       the number of .ready files in the pg_xlog/archive_status directory.
359       Usually, these values should be low, turning on the archive mechanism,
360       we usually want it to archive WAL files as fast as possible.
361
362       If the archive command fail, number of WAL in your pg_xlog directory
363       will grow until exhausting all the disk space and force PostgreSQL to
364       stop immediately.
365
366       To avoid connecting as a database superuser, a wrapper function around
367       "pg_ls_dir()" should be defined as a superuser with SECURITY DEFINER,
368       and the --lsfunc option used. This example function, if defined by a
369       superuser, will allow the script to connect as a normal user nagios
370       with --lsfunc=ls_archive_status_dir
371
372         BEGIN;
373         CREATE FUNCTION ls_archive_status_dir()
374             RETURNS SETOF TEXT
375             AS $$ SELECT pg_ls_dir('pg_xlog/archive_status') $$
376             LANGUAGE SQL
377             SECURITY DEFINER;
378         REVOKE ALL ON FUNCTION ls_archive_status_dir() FROM PUBLIC;
379         GRANT EXECUTE ON FUNCTION ls_archive_status_dir() to nagios;
380         COMMIT;
381
382       Example 1: Check that the number of ready WAL files is 10 or less on
383       host "pluto", using a wrapper function "ls_archive_status_dir" to avoid
384       the need for superuser permissions
385
386         check_postgres_archive_ready --host=pluto --critical=10 --lsfunc=ls_archive_status_dir
387
388       For MRTG output, reports the number of ready WAL files on line 1.
389
390   autovac_freeze
391       ("symlink: check_postgres_autovac_freeze") Checks how close each
392       database is to the Postgres autovacuum_freeze_max_age setting. This
393       action will only work for databases version 8.2 or higher. The
394       --warning and --critical options should be expressed as percentages.
395       The 'age' of the transactions in each database is compared to the
396       autovacuum_freeze_max_age setting (200 million by default) to generate
397       a rounded percentage. The default values are 90% for the warning and
398       95% for the critical. Databases can be filtered by use of the --include
399       and --exclude options.  See the "BASIC FILTERING" section for more
400       details.
401
402       Example 1: Give a warning when any databases on port 5432 are above 97%
403
404         check_postgres_autovac_freeze --port=5432 --warning="97%"
405
406       For MRTG output, the highest overall percentage is reported on the
407       first line, and the highest age is reported on the second line. All
408       databases which have the percentage from the first line are reported on
409       the fourth line, separated by a pipe symbol.
410
411   backends
412       ("symlink: check_postgres_backends") Checks the current number of
413       connections for one or more databases, and optionally compares it to
414       the maximum allowed, which is determined by the Postgres configuration
415       variable max_connections. The --warning and --critical options can take
416       one of three forms. First, a simple number can be given, which
417       represents the number of connections at which the alert will be given.
418       This choice does not use the max_connections setting. Second, the
419       percentage of available connections can be given. Third, a negative
420       number can be given which represents the number of connections left
421       until max_connections is reached. The default values for --warning and
422       --critical are '90%' and '95%'.  You can also filter the databases by
423       use of the --include and --exclude options.  See the "BASIC FILTERING"
424       section for more details.
425
426       To view only non-idle processes, you can use the --noidle argument.
427       Note that the user you are connecting as must be a superuser for this
428       to work properly.
429
430       Example 1: Give a warning when the number of connections on host quirm
431       reaches 120, and a critical if it reaches 150.
432
433         check_postgres_backends --host=quirm --warning=120 --critical=150
434
435       Example 2: Give a critical when we reach 75% of our max_connections
436       setting on hosts lancre or lancre2.
437
438         check_postgres_backends --warning='75%' --critical='75%' --host=lancre,lancre2
439
440       Example 3: Give a warning when there are only 10 more connection slots
441       left on host plasmid, and a critical when we have only 5 left.
442
443         check_postgres_backends --warning=-10 --critical=-5 --host=plasmid
444
445       Example 4: Check all databases except those with "test" in their name,
446       but allow ones that are named "pg_greatest". Connect as port 5432 on
447       the first two hosts, and as port 5433 on the third one. We want to
448       always throw a critical when we reach 30 or more connections.
449
450        check_postgres_backends --dbhost=hong,kong --dbhost=fooey --dbport=5432 --dbport=5433 --warning=30 --critical=30 --exclude="~test" --include="pg_greatest,~prod"
451
452       For MRTG output, the number of connections is reported on the first
453       line, and the fourth line gives the name of the database, plus the
454       current maximum_connections. If more than one database has been
455       queried, the one with the highest number of connections is output.
456
457   bloat
458       ("symlink: check_postgres_bloat") Checks the amount of bloat in tables
459       and indexes. (Bloat is generally the amount of dead unused space taken
460       up in a table or index. This space is usually reclaimed by use of the
461       VACUUM command.) This action requires that stats collection be enabled
462       on the target databases, and requires that ANALYZE is run frequently.
463       The --include and --exclude options can be used to filter out which
464       tables to look at. See the "BASIC FILTERING" section for more details.
465
466       The --warning and --critical options can be specified as sizes,
467       percents, or both.  Valid size units are bytes, kilobytes, megabytes,
468       gigabytes, terabytes, exabytes, petabytes, and zettabytes. You can
469       abbreviate all of those with the first letter. Items without units are
470       assumed to be 'bytes'. The default values are '1 GB' and '5 GB'. The
471       value represents the number of "wasted bytes", or the difference
472       between what is actually used by the table and index, and what we
473       compute that it should be.
474
475       Note that this action has two hard-coded values to avoid false alarms
476       on smaller relations. Tables must have at least 10 pages, and indexes
477       at least 15, before they can be considered by this test. If you really
478       want to adjust these values, you can look for the variables $MINPAGES
479       and $MINIPAGES at the top of the "check_bloat" subroutine. These values
480       are ignored if either --exclude or --include is used.
481
482       Only the top 10 most bloated relations are shown. You can change this
483       number by using the --perflimit option to set your own limit.
484
485       The schema named 'information_schema' is excluded from this test, as
486       the only tables it contains are small and do not change.
487
488       Please note that the values computed by this action are not precise,
489       and should be used as a guideline only. Great effort was made to
490       estimate the correct size of a table, but in the end it is only an
491       estimate. The correct index size is even more of a guess than the
492       correct table size, but both should give a rough idea of how bloated
493       things are.
494
495       Example 1: Warn if any table on port 5432 is over 100 MB bloated, and
496       critical if over 200 MB
497
498         check_postgres_bloat --port=5432 --warning='100 M' --critical='200 M'
499
500       Example 2: Give a critical if table 'orders' on host 'sami' has more
501       than 10 megs of bloat
502
503         check_postgres_bloat --host=sami --include=orders --critical='10 MB'
504
505       Example 3: Give a critical if table 'q4' on database 'sales' is over
506       50% bloated
507
508         check_postgres_bloat --db=sales --include=q4 --critical='50%'
509
510       Example 4: Give a critical any table is over 20% bloated and has over
511       150 MB of bloat:
512
513         check_postgres_bloat --port=5432 --critical='20% and 150 M'
514
515       Example 5: Give a critical any table is over 40% bloated or has over
516       500 MB of bloat:
517
518         check_postgres_bloat --port=5432 --warning='500 M or 40%'
519
520       For MRTG output, the first line gives the highest number of wasted
521       bytes for the tables, and the second line gives the highest number of
522       wasted bytes for the indexes. The fourth line gives the database name,
523       table name, and index name information. If you want to output the bloat
524       ratio instead (how many times larger the relation is compared to how
525       large it should be), just pass in "--mrtg=ratio".
526
527   checkpoint
528       ("symlink: check_postgres_checkpoint") Determines how long since the
529       last checkpoint has been run. This must run on the same server as the
530       database that is being checked (e.g. the -h flag will not work). This
531       check is meant to run on a "warm standby" server that is actively
532       processing shipped WAL files, and is meant to check that your warm
533       standby is truly 'warm'.  The data directory must be set, either by the
534       environment variable "PGDATA", or passing the "--datadir" argument. It
535       returns the number of seconds since the last checkpoint was run, as
536       determined by parsing the call to "pg_controldata". Because of this,
537       the pg_controldata executable must be available in the current path.
538       Alternatively, you can specify "PGBINDIR" as the directory that it
539       lives in.  It is also possible to use the special options --assume-prod
540       or --assume-standby-mode, if the mode found is not the one expected, a
541       CRITICAL is emitted.
542
543       At least one warning or critical argument must be set.
544
545       This action requires the Date::Parse module.
546
547       For MRTG or simple output, returns the number of seconds.
548
549   cluster_id
550       ("symlink: check_postgres_cluster-id") Checks that the Database System
551       Identifier provided by pg_controldata is the same as last time you
552       checked. This must run on the same server as the database that is being
553       checked (e.g. the -h flag will not work).  Either the --warning or the
554       --critical option should be given, but not both. The value of each one
555       is the cluster identifier, an integer value. You can run with the
556       special "--critical=0" option to find out an existing cluster
557       identifier.
558
559       Example 1: Find the initial identifier
560
561         check_postgres_cluster_id --critical=0 --datadir=/var//lib/postgresql/9.0/main
562
563       Example 2: Make sure the cluster is the same and warn if not, using the
564       result from above.
565
566         check_postgres_cluster_id  --critical=5633695740047915135
567
568       For MRTG output, returns a 1 or 0 indicating success of failure of the
569       identifier to match. A identifier must be provided as the "--mrtg"
570       argument. The fourth line always gives the current identifier.
571
572   commitratio
573       ("symlink: check_postgres_commitratio") Checks the commit ratio of all
574       databases and complains when they are too low.  There is no need to run
575       this command more than once per database cluster.  Databases can be
576       filtered with the --include and --exclude options. See the "BASIC
577       FILTERING" section for more details.  They can also be filtered by the
578       owner of the database with the --includeuser and --excludeuser options.
579       See the "USER NAME FILTERING" section for more details.
580
581       The warning and critical options should be specified as percentages.
582       There are not defaults for this action: the warning and critical must
583       be specified. The warning value cannot be greater than the critical
584       value. The output returns all databases sorted by commitratio, smallest
585       first.
586
587       Example: Warn if any database on host flagg is less than 90% in
588       commitratio, and critical if less then 80%.
589
590         check_postgres_database_commitratio --host=flagg --warning='90%' --critical='80%'
591
592       For MRTG output, returns the percentage of the database with the
593       smallest commitratio on the first line, and the name of the database on
594       the fourth line.
595
596   connection
597       ("symlink: check_postgres_connection") Simply connects, issues a
598       'SELECT version()', and leaves.  Takes no --warning or --critical
599       options.
600
601       For MRTG output, simply outputs a 1 (good connection) or a 0 (bad
602       connection) on the first line.
603
604   custom_query
605       ("symlink: check_postgres_custom_query") Runs a custom query of your
606       choosing, and parses the results.  The query itself is passed in
607       through the "query" argument, and should be kept as simple as possible.
608       If at all possible, wrap it in a view or a function to keep things
609       easier to manage. The query should return one or two columns. It is
610       required that one of the columns be named "result" and is the item that
611       will be checked against your warning and critical values. The second
612       column is for the performance data and any name can be used: this will
613       be the 'value' inside the performance data section.
614
615       At least one warning or critical argument must be specified. What these
616       are set to depends on the type of query you are running. There are four
617       types of custom_queries that can be run, specified by the "valtype"
618       argument. If none is specified, this action defaults to 'integer'. The
619       four types are:
620
621       integer: Does a simple integer comparison. The first column should be a
622       simple integer, and the warning and critical values should be the same.
623
624       string: The warning and critical are strings, and are triggered only if
625       the value in the first column matches it exactly. This is case-
626       sensitive.
627
628       time: The warning and the critical are times, and can have units of
629       seconds, minutes, hours, or days.  Each may be written singular or
630       abbreviated to just the first letter. If no units are given, seconds
631       are assumed. The first column should be an integer representing the
632       number of seconds to check.
633
634       size: The warning and the critical are sizes, and can have units of
635       bytes, kilobytes, megabytes, gigabytes, terabytes, or exabytes. Each
636       may be abbreviated to the first letter. If no units are given, bytes
637       are assumed. The first column should be an integer representing the
638       number of bytes to check.
639
640       Normally, an alert is triggered if the values returned are greater than
641       or equal to the critical or warning value. However, an option of
642       --reverse will trigger the alert if the returned value is lower than or
643       equal to the critical or warning value.
644
645       Example 1: Warn if any relation over 100 pages is named "rad", put the
646       number of pages inside the performance data section.
647
648         check_postgres_custom_query --valtype=string -w "rad" --query=
649           "SELECT relname AS result, relpages AS pages FROM pg_class WHERE relpages > 100"
650
651       Example 2: Give a critical if the "foobar" function returns a number
652       over 5MB:
653
654         check_postgres_custom_query --critical='5MB'--valtype=size --query="SELECT foobar() AS result"
655
656       Example 2: Warn if the function "snazzo" returns less than 42:
657
658         check_postgres_custom_query --critical=42 --query="SELECT snazzo() AS result" --reverse
659
660       If you come up with a useful custom_query, consider sending in a patch
661       to this program to make it into a standard action that other people can
662       use.
663
664       This action does not support MRTG or simple output yet.
665
666   database_size
667       ("symlink: check_postgres_database_size") Checks the size of all
668       databases and complains when they are too big.  There is no need to run
669       this command more than once per database cluster.  Databases can be
670       filtered with the --include and --exclude options. See the "BASIC
671       FILTERING" section for more details.  They can also be filtered by the
672       owner of the database with the --includeuser and --excludeuser options.
673       See the "USER NAME FILTERING" section for more details.
674
675       The warning and critical options can be specified as bytes, kilobytes,
676       megabytes, gigabytes, terabytes, or exabytes. Each may be abbreviated
677       to the first letter as well.  If no unit is given, the units are
678       assumed to be bytes. There are not defaults for this action: the
679       warning and critical must be specified. The warning value cannot be
680       greater than the critical value. The output returns all databases
681       sorted by size largest first, showing both raw bytes and a "pretty"
682       version of the size.
683
684       Example 1: Warn if any database on host flagg is over 1 TB in size, and
685       critical if over 1.1 TB.
686
687         check_postgres_database_size --host=flagg --warning='1 TB' --critical='1.1 t'
688
689       Example 2: Give a critical if the database template1 on port 5432 is
690       over 10 MB.
691
692         check_postgres_database_size --port=5432 --include=template1 --warning='10MB' --critical='10MB'
693
694       Example 3: Give a warning if any database on host 'tardis' owned by the
695       user 'tom' is over 5 GB
696
697         check_postgres_database_size --host=tardis --includeuser=tom --warning='5 GB' --critical='10 GB'
698
699       For MRTG output, returns the size in bytes of the largest database on
700       the first line, and the name of the database on the fourth line.
701
702   dbstats
703       ("symlink: check_postgres_dbstats") Reports information from the
704       pg_stat_database view, and outputs it in a Cacti-friendly manner. No
705       other output is supported, as the output is informational and does not
706       lend itself to alerts, such as used with Nagios. If no options are
707       given, all databases are returned, one per line. You can include a
708       specific database by use of the "--include" option, or you can use the
709       "--dbname" option.
710
711       Eleven items are returned on each line, in the format name:value,
712       separated by a single space. The items are:
713
714       backends
715           The number of currently running backends for this database.
716
717       commits
718           The total number of commits for this database since it was created
719           or reset.
720
721       rollbacks
722           The total number of rollbacks for this database since it was
723           created or reset.
724
725       read
726           The total number of disk blocks read.
727
728       hit The total number of buffer hits.
729
730       ret The total number of rows returned.
731
732       fetch
733           The total number of rows fetched.
734
735       ins The total number of rows inserted.
736
737       upd The total number of rows updated.
738
739       del The total number of rows deleted.
740
741       dbname
742           The name of the database.
743
744       Note that ret, fetch, ins, upd, and del items will always be 0 if
745       Postgres is version 8.2 or lower, as those stats were not available in
746       those versions.
747
748       If the dbname argument is given, seven additional items are returned:
749
750       idxscan
751           Total number of user index scans.
752
753       idxtupread
754           Total number of user index entries returned.
755
756       idxtupfetch
757           Total number of rows fetched by simple user index scans.
758
759       idxblksread
760           Total number of disk blocks read for all user indexes.
761
762       idxblkshit
763           Total number of buffer hits for all user indexes.
764
765       seqscan
766           Total number of sequential scans against all user tables.
767
768       seqtupread
769           Total number of tuples returned from all user tables.
770
771       Example 1: Grab the stats for a database named "products" on host
772       "willow":
773
774         check_postgres_dbstats --dbhost willow --dbname products
775
776       The output returned will be like this (all on one line, not wrapped):
777
778           backends:82 commits:58374408 rollbacks:1651 read:268435543 hit:2920381758 idxscan:310931294 idxtupread:2777040927
779           idxtupfetch:1840241349 idxblksread:62860110 idxblkshit:1107812216 seqscan:5085305 seqtupread:5370500520
780           ret:0 fetch:0 ins:0 upd:0 del:0 dbname:willow
781
782   disabled_triggers
783       ("symlink: check_postgres_disabled_triggers") Checks on the number of
784       disabled triggers inside the database.  The --warning and --critical
785       options are the number of such triggers found, and both default to "1",
786       as in normal usage having disabled triggers is a dangerous event. If
787       the database being checked is 8.3 or higher, the check is for the
788       number of triggers that are in a 'disabled' status (as opposed to being
789       'always' or 'replica'). The output will show the name of the table and
790       the name of the trigger for each disabled trigger.
791
792       Example 1: Make sure that there are no disabled triggers
793
794         check_postgres_disabled_triggers
795
796       For MRTG output, returns the number of disabled triggers on the first
797       line.
798
799   disk_space
800       ("symlink: check_postgres_disk_space") Checks on the available physical
801       disk space used by Postgres. This action requires that you have the
802       executable "/bin/df" available to report on disk sizes, and it also
803       needs to be run as a superuser, so it can examine the data_directory
804       setting inside of Postgres. The --warning and --critical options are
805       given in either sizes or percentages or both. If using sizes, the
806       standard unit types are allowed: bytes, kilobytes, gigabytes,
807       megabytes, gigabytes, terabytes, or exabytes. Each may be abbreviated
808       to the first letter only; no units at all indicates 'bytes'. The
809       default values are '90%' and '95%'.
810
811       This command checks the following things to determine all of the
812       different physical disks being used by Postgres.
813
814       data_directory - The disk that the main data directory is on.
815
816       log directory - The disk that the log files are on.
817
818       WAL file directory - The disk that the write-ahead logs are on (e.g.
819       symlinked pg_xlog or pg_wal)
820
821       tablespaces - Each tablespace that is on a separate disk.
822
823       The output shows the total size used and available on each disk, as
824       well as the percentage, ordered by highest to lowest percentage used.
825       Each item above maps to a file system: these can be included or
826       excluded. See the "BASIC FILTERING" section for more details.
827
828       Example 1: Make sure that no file system is over 90% for the database
829       on port 5432.
830
831         check_postgres_disk_space --port=5432 --warning='90%' --critical='90%'
832
833       Example 2: Check that all file systems starting with /dev/sda are
834       smaller than 10 GB and 11 GB (warning and critical)
835
836         check_postgres_disk_space --port=5432 --warning='10 GB' --critical='11 GB' --include="~^/dev/sda"
837
838       Example 4: Make sure that no file system is both over 50% and has over
839       15 GB
840
841         check_postgres_disk_space --critical='50% and 15 GB'
842
843       Example 5: Issue a warning if any file system is either over 70% full
844       or has more than 1T
845
846         check_postgres_disk_space --warning='1T or 75'
847
848       For MRTG output, returns the size in bytes of the file system on the
849       first line, and the name of the file system on the fourth line.
850
851   fsm_pages
852       ("symlink: check_postgres_fsm_pages") Checks how close a cluster is to
853       the Postgres max_fsm_pages setting.  This action will only work for
854       databases of 8.2 or higher, and it requires the contrib module
855       pg_freespacemap be installed. The --warning and --critical options
856       should be expressed as percentages. The number of used pages in the
857       free-space-map is determined by looking in the
858       pg_freespacemap_relations view, and running a formula based on the
859       formula used for outputting free-space-map pageslots in the vacuum
860       verbose command. The default values are 85% for the warning and 95% for
861       the critical.
862
863       Example 1: Give a warning when our cluster has used up 76% of the free-
864       space pageslots, with pg_freespacemap installed in database robert
865
866         check_postgres_fsm_pages --dbname=robert --warning="76%"
867
868       While you need to pass in the name of the database where
869       pg_freespacemap is installed, you only need to run this check once per
870       cluster. Also, checking this information does require obtaining special
871       locks on the free-space-map, so it is recommend you do not run this
872       check with short intervals.
873
874       For MRTG output, returns the percent of free-space-map on the first
875       line, and the number of pages currently used on the second line.
876
877   fsm_relations
878       ("symlink: check_postgres_fsm_relations") Checks how close a cluster is
879       to the Postgres max_fsm_relations setting.  This action will only work
880       for databases of 8.2 or higher, and it requires the contrib module
881       pg_freespacemap be installed. The --warning and --critical options
882       should be expressed as percentages. The number of used relations in the
883       free-space-map is determined by looking in the
884       pg_freespacemap_relations view. The default values are 85% for the
885       warning and 95% for the critical.
886
887       Example 1: Give a warning when our cluster has used up 80% of the free-
888       space relations, with pg_freespacemap installed in database dylan
889
890         check_postgres_fsm_relations --dbname=dylan --warning="75%"
891
892       While you need to pass in the name of the database where
893       pg_freespacemap is installed, you only need to run this check once per
894       cluster. Also, checking this information does require obtaining special
895       locks on the free-space-map, so it is recommend you do not run this
896       check with short intervals.
897
898       For MRTG output, returns the percent of free-space-map on the first
899       line, the number of relations currently used on the second line.
900
901   hitratio
902       ("symlink: check_postgres_hitratio") Checks the hit ratio of all
903       databases and complains when they are too low.  There is no need to run
904       this command more than once per database cluster.  Databases can be
905       filtered with the --include and --exclude options. See the "BASIC
906       FILTERING" section for more details.  They can also be filtered by the
907       owner of the database with the --includeuser and --excludeuser options.
908       See the "USER NAME FILTERING" section for more details.
909
910       The warning and critical options should be specified as percentages.
911       There are not defaults for this action: the warning and critical must
912       be specified. The warning value cannot be greater than the critical
913       value. The output returns all databases sorted by hitratio, smallest
914       first.
915
916       Example: Warn if any database on host flagg is less than 90% in
917       hitratio, and critical if less then 80%.
918
919         check_postgres_hitratio --host=flagg --warning='90%' --critical='80%'
920
921       For MRTG output, returns the percentage of the database with the
922       smallest hitratio on the first line, and the name of the database on
923       the fourth line.
924
925   hot_standby_delay
926       ("symlink: check_hot_standby_delay") Checks the streaming replication
927       lag by computing the delta between the current xlog position of a
928       master server and the replay location of a slave connected to it. The
929       slave server must be in hot_standby (e.g. read only) mode, therefore
930       the minimum version to use this action is Postgres 9.0. The --warning
931       and --critical options are the delta between the xlog locations. Since
932       these values are byte offsets in the WAL they should match the expected
933       transaction volume of your application to prevent false positives or
934       negatives.
935
936       The first "--dbname", "--host", and "--port", etc. options are
937       considered the master; the second belongs to the slave.
938
939       Byte values should be based on the volume of transactions needed to
940       have the streaming replication disconnect from the master because of
941       too much lag, determined by the Postgres configuration variable
942       wal_keep_segments.  For units of time, valid units are 'seconds',
943       'minutes', 'hours', or 'days'.  Each may be written singular or
944       abbreviated to just the first letter. When specifying both, in the form
945       'bytes and time', both conditions must be true for the threshold to be
946       met.
947
948       You must provide information on how to reach the databases by providing
949       a comma separated list to the --dbhost and --dbport parameters, such as
950       "--dbport=5432,5543". If not given, the action fails.
951
952       Example 1: Warn a database with a local replica on port 5433 is behind
953       on any xlog replay at all
954
955         check_hot_standby_delay --dbport=5432,5433 --warning='1'
956
957       Example 2: Give a critical if the last transaction replica1 receives is
958       more than 10 minutes ago
959
960         check_hot_standby_delay --dbhost=master,replica1 --critical='10 min'
961
962       Example 3: Allow replica1 to be 1 WAL segment behind, if the master is
963       momentarily seeing more activity than the streaming replication
964       connection can handle, or 10 minutes behind, if the master is seeing
965       very little activity and not processing any transactions, but not both,
966       which would indicate a lasting problem with the replication connection.
967
968         check_hot_standby_delay --dbhost=master,replica1 --warning='1048576 and 2 min' --critical='16777216 and 10 min'
969
970   relation_size
971   index_size
972   table_size
973   indexes_size
974   total_relation_size
975       (symlinks: "check_postgres_relation_size", "check_postgres_index_size",
976       "check_postgres_table_size", "check_postgres_indexes_size", and
977       "check_postgres_total_relation_size")
978
979       The actions relation_size and index_size check for a relation (table,
980       index, materialized view), respectively an index that has grown too
981       big, using the pg_relation_size() function.
982
983       The action table_size checks tables and materialized views using
984       pg_table_size(), i.e. including relation forks and TOAST table.
985
986       The action indexes_size checks tables and materialized views for the
987       size of the attached indexes using pg_indexes_size().
988
989       The action total_relation_size checks relations using
990       pg_total_relation_size(), i.e. including relation forks, indexes and
991       TOAST table.
992
993       Relations can be filtered with the --include and --exclude options. See
994       the "BASIC FILTERING" section for more details. Relations can also be
995       filtered by the user that owns them, by using the --includeuser and
996       --excludeuser options.  See the "USER NAME FILTERING" section for more
997       details.
998
999       The values for the --warning and --critical options are file sizes, and
1000       may have units of bytes, kilobytes, megabytes, gigabytes, terabytes, or
1001       exabytes.  Each can be abbreviated to the first letter. If no units are
1002       given, bytes are assumed. There are no default values: both the warning
1003       and the critical option must be given. The return text shows the size
1004       of the largest relation found.
1005
1006       If the --showperf option is enabled, all of the relations with their
1007       sizes will be given. To prevent this, it is recommended that you set
1008       the --perflimit option, which will cause the query to do a "ORDER BY
1009       size DESC LIMIT (perflimit)".
1010
1011       Example 1: Give a critical if any table is larger than 600MB on host
1012       burrick.
1013
1014         check_postgres_table_size --critical='600 MB' --warning='600 MB' --host=burrick
1015
1016       Example 2: Warn if the table products is over 4 GB in size, and give a
1017       critical at 4.5 GB.
1018
1019         check_postgres_table_size --host=burrick --warning='4 GB' --critical='4.5 GB' --include=products
1020
1021       Example 3: Warn if any index not owned by postgres goes over 500 MB.
1022
1023         check_postgres_index_size --port=5432 --excludeuser=postgres -w 500MB -c 600MB
1024
1025       For MRTG output, returns the size in bytes of the largest relation, and
1026       the name of the database and relation as the fourth line.
1027
1028   last_analyze
1029   last_vacuum
1030   last_autoanalyze
1031   last_autovacuum
1032       (symlinks: "check_postgres_last_analyze", "check_postgres_last_vacuum",
1033       "check_postgres_last_autoanalyze", and
1034       "check_postgres_last_autovacuum") Checks how long it has been since
1035       vacuum (or analyze) was last run on each table in one or more
1036       databases. Use of these actions requires that the target database is
1037       version 8.3 or greater, or that the version is 8.2 and the
1038       configuration variable stats_row_level has been enabled. Tables can be
1039       filtered with the --include and --exclude options. See the "BASIC
1040       FILTERING" section for more details.  Tables can also be filtered by
1041       their owner by use of the --includeuser and --excludeuser options.  See
1042       the "USER NAME FILTERING" section for more details.
1043
1044       The units for --warning and --critical are specified as times.  Valid
1045       units are seconds, minutes, hours, and days; all can be abbreviated to
1046       the first letter. If no units are given, 'seconds' are assumed. The
1047       default values are '1 day' and '2 days'. Please note that there are
1048       cases in which this field does not get automatically populated. If
1049       certain tables are giving you problems, make sure that they have dead
1050       rows to vacuum, or just exclude them from the test.
1051
1052       The schema named 'information_schema' is excluded from this test, as
1053       the only tables it contains are small and do not change.
1054
1055       Note that the non-'auto' versions will also check on the auto versions
1056       as well. In other words, using last_vacuum will report on the last
1057       vacuum, whether it was a normal vacuum, or one run by the autovacuum
1058       daemon.
1059
1060       Example 1: Warn if any table has not been vacuumed in 3 days, and give
1061       a critical at a week, for host wormwood
1062
1063         check_postgres_last_vacuum --host=wormwood --warning='3d' --critical='7d'
1064
1065       Example 2: Same as above, but skip tables belonging to the users 'eve'
1066       or 'mallory'
1067
1068         check_postgres_last_vacuum --host=wormwood --warning='3d' --critical='7d' --excludeuser=eve,mallory
1069
1070       For MRTG output, returns (on the first line) the LEAST amount of time
1071       in seconds since a table was last vacuumed or analyzed. The fourth line
1072       returns the name of the database and name of the table.
1073
1074   listener
1075       ("symlink: check_postgres_listener") Confirm that someone is listening
1076       for one or more specific strings (using the LISTEN/NOTIFY system), by
1077       looking at the pg_listener table.  Only one of warning or critical is
1078       needed. The format is a simple string representing the LISTEN target,
1079       or a tilde character followed by a string for a regular expression
1080       check.  Note that this check will not work on versions of Postgres 9.0
1081       or higher.
1082
1083       Example 1: Give a warning if nobody is listening for the string
1084       bucardo_mcp_ping on ports 5555 and 5556
1085
1086         check_postgres_listener --port=5555,5556 --warning=bucardo_mcp_ping
1087
1088       Example 2: Give a critical if there are no active LISTEN requests
1089       matching 'grimm' on database oskar
1090
1091         check_postgres_listener --db oskar --critical=~grimm
1092
1093       For MRTG output, returns a 1 or a 0 on the first, indicating success or
1094       failure. The name of the notice must be provided via the --mrtg option.
1095
1096   locks
1097       ("symlink: check_postgres_locks") Check the total number of locks on
1098       one or more databases. There is no need to run this more than once per
1099       database cluster. Databases can be filtered with the --include and
1100       --exclude options. See the "BASIC FILTERING" section for more details.
1101
1102       The --warning and --critical options can be specified as simple
1103       numbers, which represent the total number of locks, or they can be
1104       broken down by type of lock.  Valid lock names are 'total', 'waiting',
1105       or the name of a lock type used by Postgres.  These names are case-
1106       insensitive and do not need the "lock" part on the end, so exclusive
1107       will match 'ExclusiveLock'. The format is name=number, with different
1108       items separated by colons or semicolons (or any other symbol).
1109
1110       Example 1: Warn if the number of locks is 100 or more, and critical if
1111       200 or more, on host garrett
1112
1113         check_postgres_locks --host=garrett --warning=100 --critical=200
1114
1115       Example 2: On the host artemus, warn if 200 or more locks exist, and
1116       give a critical if over 250 total locks exist, or if over 20 exclusive
1117       locks exist, or if over 5 connections are waiting for a lock.
1118
1119         check_postgres_locks --host=artemus --warning=200 --critical="total=250:waiting=5:exclusive=20"
1120
1121       For MRTG output, returns the number of locks on the first line, and the
1122       name of the database on the fourth line.
1123
1124   logfile
1125       ("symlink: check_postgres_logfile") Ensures that the logfile is in the
1126       expected location and is being logged to.  This action issues a command
1127       that throws an error on each database it is checking, and ensures that
1128       the message shows up in the logs. It scans the various log_* settings
1129       inside of Postgres to figure out where the logs should be.  If you are
1130       using syslog, it does a rough (but not foolproof) scan of
1131       /etc/syslog.conf. Alternatively, you can provide the name of the
1132       logfile with the --logfile option. This is especially useful if the
1133       logs have a custom rotation scheme driven be an external program. The
1134       --logfile option supports the following escape characters: "%Y %m %d
1135       %H", which represent the current year, month, date, and hour
1136       respectively. An error is always reported as critical unless the
1137       warning option has been passed in as a non-zero value. Other than that
1138       specific usage, the "--warning" and "--critical" options should not be
1139       used.
1140
1141       Example 1: On port 5432, ensure the logfile is being written to the
1142       file /home/greg/pg8.2.log
1143
1144         check_postgres_logfile --port=5432 --logfile=/home/greg/pg8.2.log
1145
1146       Example 2: Same as above, but raise a warning, not a critical
1147
1148         check_postgres_logfile --port=5432 --logfile=/home/greg/pg8.2.log -w 1
1149
1150       For MRTG output, returns a 1 or 0 on the first line, indicating success
1151       or failure. In case of a failure, the fourth line will provide more
1152       detail on the failure encountered.
1153
1154   new_version_bc
1155       ("symlink: check_postgres_new_version_bc") Checks if a newer version of
1156       the Bucardo program is available. The current version is obtained by
1157       running "bucardo_ctl --version".  If a major upgrade is available, a
1158       warning is returned. If a revision upgrade is available, a critical is
1159       returned. (Bucardo is a master to slave, and master to master
1160       replication system for Postgres: see https://bucardo.org/ for more
1161       information).  See also the information on the "--get_method" option.
1162
1163   new_version_box
1164       ("symlink: check_postgres_new_version_box") Checks if a newer version
1165       of the boxinfo program is available. The current version is obtained by
1166       running "boxinfo.pl --version".  If a major upgrade is available, a
1167       warning is returned. If a revision upgrade is available, a critical is
1168       returned. (boxinfo is a program for grabbing important information from
1169       a server and putting it into a HTML format: see
1170       https://bucardo.org/Boxinfo/ for more information). See also the
1171       information on the "--get_method" option.
1172
1173   new_version_cp
1174       ("symlink: check_postgres_new_version_cp") Checks if a newer version of
1175       this program (check_postgres.pl) is available, by grabbing the version
1176       from a small text file on the main page of the home page for the
1177       project. Returns a warning if the returned version does not match the
1178       one you are running. Recommended interval to check is once a day. See
1179       also the information on the "--get_method" option.
1180
1181   new_version_pg
1182       ("symlink: check_postgres_new_version_pg") Checks if a newer revision
1183       of Postgres exists for each database connected to. Note that this only
1184       checks for revision, e.g.  going from 8.3.6 to 8.3.7. Revisions are
1185       always 100% binary compatible and involve no dump and restore to
1186       upgrade. Revisions are made to address bugs, so upgrading as soon as
1187       possible is always recommended. Returns a warning if you do not have
1188       the latest revision.  It is recommended this check is run at least once
1189       a day. See also the information on the "--get_method" option.
1190
1191   new_version_tnm
1192       ("symlink: check_postgres_new_version_tnm") Checks if a newer version
1193       of the tail_n_mail program is available. The current version is
1194       obtained by running "tail_n_mail --version". If a major upgrade is
1195       available, a warning is returned. If a revision upgrade is available, a
1196       critical is returned. (tail_n_mail is a log monitoring tool that can
1197       send mail when interesting events appear in your Postgres logs.  See:
1198       https://bucardo.org/tail_n_mail/ for more information).  See also the
1199       information on the "--get_method" option.
1200
1201   pgb_pool_cl_active
1202   pgb_pool_cl_waiting
1203   pgb_pool_sv_active
1204   pgb_pool_sv_idle
1205   pgb_pool_sv_used
1206   pgb_pool_sv_tested
1207   pgb_pool_sv_login
1208   pgb_pool_maxwait
1209       (symlinks: "check_postgres_pgb_pool_cl_active",
1210       "check_postgres_pgb_pool_cl_waiting",
1211       "check_postgres_pgb_pool_sv_active", "check_postgres_pgb_pool_sv_idle",
1212       "check_postgres_pgb_pool_sv_used", "check_postgres_pgb_pool_sv_tested",
1213       "check_postgres_pgb_pool_sv_login", and
1214       "check_postgres_pgb_pool_maxwait")
1215
1216       Examines pgbouncer's pool statistics. Each pool has a set of "client"
1217       connections, referring to connections from external clients, and
1218       "server" connections, referring to connections to PostgreSQL itself.
1219       The related check_postgres actions are prefixed by "cl_" and "sv_",
1220       respectively. Active client connections are those connections currently
1221       linked with an active server connection. Client connections may also be
1222       "waiting", meaning they have not yet been allocated a server
1223       connection. Server connections are "active" (linked to a client),
1224       "idle" (standing by for a client connection to link with), "used" (just
1225       unlinked from a client, and not yet returned to the idle pool),
1226       "tested" (currently being tested) and "login" (in the process of
1227       logging in). The maxwait value shows how long in seconds the oldest
1228       waiting client connection has been waiting.
1229
1230   pgbouncer_backends
1231       ("symlink: check_postgres_pgbouncer_backends") Checks the current
1232       number of connections for one or more databases through pgbouncer, and
1233       optionally compares it to the maximum allowed, which is determined by
1234       the pgbouncer configuration variable max_client_conn. The --warning and
1235       --critical options can take one of three forms. First, a simple number
1236       can be given, which represents the number of connections at which the
1237       alert will be given.  This choice does not use the max_connections
1238       setting. Second, the percentage of available connections can be given.
1239       Third, a negative number can be given which represents the number of
1240       connections left until max_connections is reached. The default values
1241       for --warning and --critical are '90%' and '95%'.  You can also filter
1242       the databases by use of the --include and --exclude options.  See the
1243       "BASIC FILTERING" section for more details.
1244
1245       To view only non-idle processes, you can use the --noidle argument.
1246       Note that the user you are connecting as must be a superuser for this
1247       to work properly.
1248
1249       Example 1: Give a warning when the number of connections on host quirm
1250       reaches 120, and a critical if it reaches 150.
1251
1252         check_postgres_pgbouncer_backends --host=quirm --warning=120 --critical=150 -p 6432 -u pgbouncer
1253
1254       Example 2: Give a critical when we reach 75% of our max_connections
1255       setting on hosts lancre or lancre2.
1256
1257         check_postgres_pgbouncer_backends --warning='75%' --critical='75%' --host=lancre,lancre2 -p 6432 -u pgbouncer
1258
1259       Example 3: Give a warning when there are only 10 more connection slots
1260       left on host plasmid, and a critical when we have only 5 left.
1261
1262         check_postgres_pgbouncer_backends --warning=-10 --critical=-5 --host=plasmid -p 6432 -u pgbouncer
1263
1264       For MRTG output, the number of connections is reported on the first
1265       line, and the fourth line gives the name of the database, plus the
1266       current max_client_conn. If more than one database has been queried,
1267       the one with the highest number of connections is output.
1268
1269   pgbouncer_checksum
1270       ("symlink: check_postgres_pgbouncer_checksum") Checks that all the
1271       pgBouncer settings are the same as last time you checked.  This is done
1272       by generating a checksum of a sorted list of setting names and their
1273       values. Note that you shouldn't specify the database name, it will
1274       automatically default to pgbouncer.  Either the --warning or the
1275       --critical option should be given, but not both. The value of each one
1276       is the checksum, a 32-character hexadecimal value. You can run with the
1277       special "--critical=0" option to find out an existing checksum.
1278
1279       This action requires the Digest::MD5 module.
1280
1281       Example 1: Find the initial checksum for pgbouncer configuration on
1282       port 6432 using the default user (usually postgres)
1283
1284         check_postgres_pgbouncer_checksum --port=6432 --critical=0
1285
1286       Example 2: Make sure no settings have changed and warn if so, using the
1287       checksum from above.
1288
1289         check_postgres_pgbouncer_checksum --port=6432 --warning=cd2f3b5e129dc2b4f5c0f6d8d2e64231
1290
1291       For MRTG output, returns a 1 or 0 indicating success of failure of the
1292       checksum to match. A checksum must be provided as the "--mrtg"
1293       argument. The fourth line always gives the current checksum.
1294
1295   pgagent_jobs
1296       ("symlink: check_postgres_pgagent_jobs") Checks that all the pgAgent
1297       jobs that have executed in the preceding interval of time have
1298       succeeded. This is done by checking for any steps that have a non-zero
1299       result.
1300
1301       Either "--warning" or "--critical", or both, may be specified as times,
1302       and jobs will be checked for failures withing the specified periods of
1303       time before the current time. Valid units are seconds, minutes, hours,
1304       and days; all can be abbreviated to the first letter. If no units are
1305       given, 'seconds' are assumed.
1306
1307       Example 1: Give a critical when any jobs executed in the last day have
1308       failed.
1309
1310         check_postgres_pgagent_jobs --critical=1d
1311
1312       Example 2: Give a warning when any jobs executed in the last week have
1313       failed.
1314
1315         check_postgres_pgagent_jobs --warning=7d
1316
1317       Example 3: Give a critical for jobs that have failed in the last 2
1318       hours and a warning for jobs that have failed in the last 4 hours:
1319
1320         check_postgres_pgagent_jobs --critical=2h --warning=4h
1321
1322   prepared_txns
1323       ("symlink: check_postgres_prepared_txns") Check on the age of any
1324       existing prepared transactions.  Note that most people will NOT use
1325       prepared transactions, as they are part of two-part commit and
1326       complicated to maintain. They should also not be confused with prepared
1327       STATEMENTS, which is what most people think of when they hear prepare.
1328       The default value for a warning is 1 second, to detect any use of
1329       prepared transactions, which is probably a mistake on most systems.
1330       Warning and critical are the number of seconds a prepared transaction
1331       has been open before an alert is given.
1332
1333       Example 1: Give a warning on detecting any prepared transactions:
1334
1335         check_postgres_prepared_txns -w 0
1336
1337       Example 2: Give a critical if any prepared transaction has been open
1338       longer than 10 seconds, but allow up to 360 seconds for the database
1339       'shrike':
1340
1341         check_postgres_prepared_txns --critical=10 --exclude=shrike
1342         check_postgres_prepared_txns --critical=360 --include=shrike
1343
1344       For MRTG output, returns the number of seconds the oldest transaction
1345       has been open as the first line, and which database is came from as the
1346       final line.
1347
1348   query_runtime
1349       ("symlink: check_postgres_query_runtime") Checks how long a specific
1350       query takes to run, by executing a "EXPLAIN ANALYZE" against it. The
1351       --warning and --critical options are the maximum amount of time the
1352       query should take. Valid units are seconds, minutes, and hours; any can
1353       be abbreviated to the first letter. If no units are given, 'seconds'
1354       are assumed.  Both the warning and the critical option must be given.
1355       The name of the view or function to be run must be passed in to the
1356       --queryname option. It must consist of a single word (or schema.word),
1357       with optional parens at the end.
1358
1359       Example 1: Give a critical if the function named "speedtest" fails to
1360       run in 10 seconds or less.
1361
1362         check_postgres_query_runtime --queryname='speedtest()' --critical=10 --warning=10
1363
1364       For MRTG output, reports the time in seconds for the query to complete
1365       on the first line. The fourth line lists the database.
1366
1367   query_time
1368       ("symlink: check_postgres_query_time") Checks the length of running
1369       queries on one or more databases.  There is no need to run this more
1370       than once on the same database cluster. Note that this already excludes
1371       queries that are "idle in transaction". Databases can be filtered by
1372       using the --include and --exclude options. See the "BASIC FILTERING"
1373       section for more details. You can also filter on the user running the
1374       query with the --includeuser and --excludeuser options.  See the "USER
1375       NAME FILTERING" section for more details.
1376
1377       The values for the --warning and --critical options are amounts of
1378       time, and at least one must be provided (no defaults). Valid units are
1379       'seconds', 'minutes', 'hours', or 'days'. Each may be written singular
1380       or abbreviated to just the first letter. If no units are given, the
1381       unit is assumed to be seconds.
1382
1383       This action requires Postgres 8.1 or better.
1384
1385       Example 1: Give a warning if any query has been running longer than 3
1386       minutes, and a critical if longer than 5 minutes.
1387
1388         check_postgres_query_time --port=5432 --warning='3 minutes' --critical='5 minutes'
1389
1390       Example 2: Using default values (2 and 5 minutes), check all databases
1391       except those starting with 'template'.
1392
1393         check_postgres_query_time --port=5432 --exclude=~^template
1394
1395       Example 3: Warn if user 'don' has a query running over 20 seconds
1396
1397         check_postgres_query_time --port=5432 --includeuser=don --warning=20s
1398
1399       For MRTG output, returns the length in seconds of the longest running
1400       query on the first line. The fourth line gives the name of the
1401       database.
1402
1403   replicate_row
1404       ("symlink: check_postgres_replicate_row") Checks that master-slave
1405       replication is working to one or more slaves.
1406
1407       The first "--dbname", "--host", and "--port", etc. options are
1408       considered the master; subsequent uses are the slaves.  The values or
1409       the --warning and --critical options are units of time, and at least
1410       one must be provided (no defaults). Valid units are 'seconds',
1411       'minutes', 'hours', or 'days'. Each may be written singular or
1412       abbreviated to just the first letter.  If no units are given, the units
1413       are assumed to be seconds.
1414
1415       This check updates a single row on the master, and then measures how
1416       long it takes to be applied to the slaves. To do this, you need to pick
1417       a table that is being replicated, then find a row that can be changed,
1418       and is not going to be changed by any other process. A specific column
1419       of this row will be changed from one value to another. All of this is
1420       fed to the "repinfo" option, and should contain the following options,
1421       separated by commas: table name, primary key, key id, column, first
1422       value, second value.
1423
1424       Example 1: Slony is replicating a table named 'orders' from host
1425       'alpha' to host 'beta', in the database 'sales'. The primary key of the
1426       table is named id, and we are going to test the row with an id of 3
1427       (which is historical and never changed). There is a column named
1428       'salesrep' that we are going to toggle from a value of 'slon' to 'nols'
1429       to check on the replication. We want to throw a warning if the
1430       replication does not happen within 10 seconds.
1431
1432         check_postgres_replicate_row --host=alpha --dbname=sales --host=beta
1433         --dbname=sales --warning=10 --repinfo=orders,id,3,salesrep,slon,nols
1434
1435       Example 2: Bucardo is replicating a table named 'receipt' from host
1436       'green' to hosts 'red', 'blue', and 'yellow'. The database for both
1437       sides is 'public'.  The slave databases are running on port 5455. The
1438       primary key is named 'receipt_id', the row we want to use has a value
1439       of 9, and the column we want to change for the test is called 'zone'.
1440       We'll toggle between 'north' and 'south' for the value of this column,
1441       and throw a critical if the change is not on all three slaves within 5
1442       seconds.
1443
1444        check_postgres_replicate_row --host=green --port=5455 --host=red,blue,yellow
1445         --critical=5 --repinfo=receipt,receipt_id,9,zone,north,south
1446
1447       For MRTG output, returns on the first line the time in seconds the
1448       replication takes to finish.  The maximum time is set to 4 minutes 30
1449       seconds: if no replication has taken place in that long a time, an
1450       error is thrown.
1451
1452   replication_slots
1453       ("symlink: check_postgres_replication_slots")  Check the quantity of
1454       WAL retained for any replication slots in the target database cluster.
1455       This is handy for monitoring environments where all WAL archiving and
1456       replication is taking place over replication slots.
1457
1458       Warning and critical are total bytes retained for the slot. E.g:
1459
1460         check_postgres_replication_slots --port=5432 --host=yellow -warning=32M -critical=64M
1461
1462       Specific named slots can be monitored using --include/--exclude
1463
1464   same_schema
1465       ("symlink: check_postgres_same_schema") Verifies that two or more
1466       databases are identical as far as their schema (but not the data
1467       within). This is particularly handy for making sure your slaves have
1468       not been modified or corrupted in any way when using master to slave
1469       replication. Unlike most other actions, this has no warning or critical
1470       criteria - the databases are either in sync, or are not.  If they are
1471       different, a detailed list of the differences is presented.
1472
1473       You may want to exclude or filter out certain differences. The way to
1474       do this is to add strings to the "--filter" option. To exclude a type
1475       of object, use "noname", where 'name' is the type of object, for
1476       example, "noschema". To exclude objects of a certain type by a regular
1477       expression against their name, use "noname=regex". See the examples
1478       below for a better understanding.
1479
1480       The types of objects that can be filtered include:
1481
1482       user
1483       schema
1484       table
1485       view
1486       index
1487       sequence
1488       constraint
1489       trigger
1490       function
1491
1492       The filter option "noposition"  prevents verification of the position
1493       of columns within a table.
1494
1495       The filter option "nofuncbody" prevents comparison of the bodies of all
1496       functions.
1497
1498       The filter option "noperm" prevents comparison of object permissions.
1499
1500       To provide the second database, just append the differences to the
1501       first one by a call to the appropriate connection argument. For
1502       example, to compare databases on hosts alpha and bravo, use
1503       "--dbhost=alpha,bravo". Also see the examples below.
1504
1505       If only a single host is given, it is assumed we are doing a "time-
1506       based" report.  The first time this is run a snapshot of all the items
1507       in the database is saved to a local file. When you run it again, that
1508       snapshot is read in and becomes "database #2" and is compared to the
1509       current database.
1510
1511       To replace the old stored file with the new version, use the --replace
1512       argument.
1513
1514       If you need to write the stored file to a specific directory, use the
1515       --audit-file-dir argument.
1516
1517       To avoid false positives on value based checks caused by replication
1518       lag on asynchronous replicas, use the --assume-async option.
1519
1520       To enable snapshots at various points in time, you can use the
1521       "--suffix" argument to make the filenames unique to each run. See the
1522       examples below.
1523
1524       Example 1: Verify that two databases on hosts star and line are the
1525       same:
1526
1527         check_postgres_same_schema --dbhost=star,line
1528
1529       Example 2: Same as before, but exclude any triggers with "slony" in
1530       their name
1531
1532         check_postgres_same_schema --dbhost=star,line --filter="notrigger=slony"
1533
1534       Example 3: Same as before, but also exclude all indexes
1535
1536         check_postgres_same_schema --dbhost=star,line --filter="notrigger=slony noindexes"
1537
1538       Example 4: Check differences for the database "battlestar" on different
1539       ports
1540
1541         check_postgres_same_schema --dbname=battlestar --dbport=5432,5544
1542
1543       Example 5: Create a daily and weekly snapshot file
1544
1545         check_postgres_same_schema --dbname=cylon --suffix=daily
1546         check_postgres_same_schema --dbname=cylon --suffix=weekly
1547
1548       Example 6: Run a historical comparison, then replace the file
1549
1550         check_postgres_same_schema --dbname=cylon --suffix=daily --replace
1551
1552       Example 7: Verify that two databases on hosts star and line are the
1553       same, excluding value data (i.e. sequence last_val):
1554
1555         check_postgres_same_schema --dbhost=star,line --assume-async
1556
1557   sequence
1558       ("symlink: check_postgres_sequence") Checks how much room is left on
1559       all sequences in the database.  This is measured as the percent of
1560       total possible values that have been used for each sequence.  The
1561       --warning and --critical options should be expressed as percentages.
1562       The default values are 85% for the warning and 95% for the critical.
1563       You may use --include and --exclude to control which sequences are to
1564       be checked. Note that this check does account for unusual minvalue and
1565       increment by values. By default it does not care if the sequence is set
1566       to cycle or not, and by passing --skipcycled sequenced set to cycle are
1567       reported with 0% usage.
1568
1569       The output for Nagios gives the name of the sequence, the percentage
1570       used, and the number of 'calls' left, indicating how many more times
1571       nextval can be called on that sequence before running into the maximum
1572       value.
1573
1574       The output for MRTG returns the highest percentage across all sequences
1575       on the first line, and the name of each sequence with that percentage
1576       on the fourth line, separated by a "|" (pipe) if there are more than
1577       one sequence at that percentage.
1578
1579       Example 1: Give a warning if any sequences are approaching 95% full.
1580
1581         check_postgres_sequence --dbport=5432 --warning=95%
1582
1583       Example 2: Check that the sequence named "orders_id_seq" is not more
1584       than half full.
1585
1586         check_postgres_sequence --dbport=5432 --critical=50% --include=orders_id_seq
1587
1588   settings_checksum
1589       ("symlink: check_postgres_settings_checksum") Checks that all the
1590       Postgres settings are the same as last time you checked.  This is done
1591       by generating a checksum of a sorted list of setting names and their
1592       values. Note that different users in the same database may have
1593       different checksums, due to ALTER USER usage, and due to the fact that
1594       superusers see more settings than ordinary users. Either the --warning
1595       or the --critical option should be given, but not both. The value of
1596       each one is the checksum, a 32-character hexadecimal value. You can run
1597       with the special "--critical=0" option to find out an existing
1598       checksum.
1599
1600       This action requires the Digest::MD5 module.
1601
1602       Example 1: Find the initial checksum for the database on port 5555
1603       using the default user (usually postgres)
1604
1605         check_postgres_settings_checksum --port=5555 --critical=0
1606
1607       Example 2: Make sure no settings have changed and warn if so, using the
1608       checksum from above.
1609
1610         check_postgres_settings_checksum --port=5555 --warning=cd2f3b5e129dc2b4f5c0f6d8d2e64231
1611
1612       For MRTG output, returns a 1 or 0 indicating success of failure of the
1613       checksum to match. A checksum must be provided as the "--mrtg"
1614       argument. The fourth line always gives the current checksum.
1615
1616   slony_status
1617       ("symlink: check_postgres_slony_status") Checks in the status of a
1618       Slony cluster by looking at the results of Slony's sl_status view. This
1619       is returned as the number of seconds of "lag time".  The --warning and
1620       --critical options should be expressed as times. The default values are
1621       60 seconds for the warning and 300 seconds for the critical.
1622
1623       The optional argument --schema indicated the schema that Slony is
1624       installed under. If it is not given, the schema will be determined
1625       automatically each time this check is run.
1626
1627       Example 1: Give a warning if any Slony is lagged by more than 20
1628       seconds
1629
1630         check_postgres_slony_status --warning 20
1631
1632       Example 2: Give a critical if Slony, installed under the schema
1633       "_slony", is over 10 minutes lagged
1634
1635         check_postgres_slony_status --schema=_slony --critical=600
1636
1637   timesync
1638       ("symlink: check_postgres_timesync") Compares the local system time
1639       with the time reported by one or more databases.  The --warning and
1640       --critical options represent the number of seconds between the two
1641       systems before an alert is given. If neither is specified, the default
1642       values are used, which are '2' and '5'. The warning value cannot be
1643       greater than the critical value. Due to the non-exact nature of this
1644       test, values of '0' or '1' are not recommended.
1645
1646       The string returned shows the time difference as well as the time on
1647       each side written out.
1648
1649       Example 1: Check that databases on hosts ankh, morpork, and klatch are
1650       no more than 3 seconds off from the local time:
1651
1652         check_postgres_timesync --host=ankh,morpork,klatch --critical=3
1653
1654       For MRTG output, returns one the first line the number of seconds
1655       difference between the local time and the database time. The fourth
1656       line returns the name of the database.
1657
1658   txn_idle
1659       ("symlink: check_postgres_txn_idle") Checks the number and duration of
1660       "idle in transaction" queries on one or more databases. There is no
1661       need to run this more than once on the same database cluster. Databases
1662       can be filtered by using the --include and --exclude options. See the
1663       "BASIC FILTERING" section below for more details.
1664
1665       The --warning and --critical options are given as units of time, signed
1666       integers, or integers for units of time, and at least one must be
1667       provided (there are no defaults). Valid units are 'seconds', 'minutes',
1668       'hours', or 'days'. Each may be written singular or abbreviated to just
1669       the first letter. If no units are given and the numbers are unsigned,
1670       the units are assumed to be seconds.
1671
1672       This action requires Postgres 8.3 or better.
1673
1674       As of PostgreSQL 10, you can just GRANT pg_read_all_stats to an
1675       unprivileged user account.  In all earlier versions, superuser
1676       privileges are required to see the queries of all users in the system;
1677       UNKNOWN is returned if queries cannot be checked. To only include
1678       queries by the connecting user, use --includeuser.
1679
1680       Example 1: Give a warning if any connection has been idle in
1681       transaction for more than 15 seconds:
1682
1683         check_postgres_txn_idle --port=5432 --warning='15 seconds'
1684
1685       Example 2: Give a warning if there are 50 or more transactions
1686
1687         check_postgres_txn_idle --port=5432 --warning='+50'
1688
1689       Example 3: Give a critical if 5 or more connections have been idle in
1690       transaction for more than 10 seconds:
1691
1692         check_postgres_txn_idle --port=5432 --critical='5 for 10 seconds'
1693
1694       For MRTG output, returns the time in seconds the longest idle
1695       transaction has been running. The fourth line returns the name of the
1696       database and other information about the longest transaction.
1697
1698   txn_time
1699       ("symlink: check_postgres_txn_time") Checks the length of open
1700       transactions on one or more databases.  There is no need to run this
1701       command more than once per database cluster.  Databases can be filtered
1702       by use of the --include and --exclude options. See the "BASIC
1703       FILTERING" section for more details. The owner of the transaction can
1704       also be filtered, by use of the --includeuser and --excludeuser
1705       options.  See the "USER NAME FILTERING" section for more details.
1706
1707       The values or the --warning and --critical options are units of time,
1708       and at least one must be provided (no default). Valid units are
1709       'seconds', 'minutes', 'hours', or 'days'. Each may be written singular
1710       or abbreviated to just the first letter.  If no units are given, the
1711       units are assumed to be seconds.
1712
1713       This action requires Postgres 8.3 or better.
1714
1715       Example 1: Give a critical if any transaction has been open for more
1716       than 10 minutes:
1717
1718         check_postgres_txn_time --port=5432 --critical='10 minutes'
1719
1720       Example 1: Warn if user 'warehouse' has a transaction open over 30
1721       seconds
1722
1723         check_postgres_txn_time --port-5432 --warning=30s --includeuser=warehouse
1724
1725       For MRTG output, returns the maximum time in seconds a transaction has
1726       been open on the first line. The fourth line gives the name of the
1727       database.
1728
1729   txn_wraparound
1730       ("symlink: check_postgres_txn_wraparound") Checks how close to
1731       transaction wraparound one or more databases are getting.  The
1732       --warning and --critical options indicate the number of transactions
1733       done, and must be a positive integer.  If either option is not given,
1734       the default values of 1.3 and 1.4 billion are used. There is no need to
1735       run this command more than once per database cluster. For a more
1736       detailed discussion of what this number represents and what to do about
1737       it, please visit the page
1738       <https://www.postgresql.org/docs/current/static/routine-vacuuming.html#VACUUM-FOR-WRAPAROUND>
1739
1740       The warning and critical values can have underscores in the number for
1741       legibility, as Perl does.
1742
1743       Example 1: Check the default values for the localhost database
1744
1745         check_postgres_txn_wraparound --host=localhost
1746
1747       Example 2: Check port 6000 and give a critical when 1.7 billion
1748       transactions are hit:
1749
1750         check_postgres_txn_wraparound --port=6000 --critical=1_700_000_000
1751
1752       For MRTG output, returns the highest number of transactions for all
1753       databases on line one, while line 4 indicates which database it is.
1754
1755   version
1756       ("symlink: check_postgres_version") Checks that the required version of
1757       Postgres is running. The --warning and --critical options (only one is
1758       required) must be of the format X.Y or X.Y.Z where X is the major
1759       version number, Y is the minor version number, and Z is the revision.
1760
1761       Example 1: Give a warning if the database on port 5678 is not version
1762       8.4.10:
1763
1764         check_postgres_version --port=5678 -w=8.4.10
1765
1766       Example 2: Give a warning if any databases on hosts valley,grain, or
1767       sunshine is not 8.3:
1768
1769         check_postgres_version -H valley,grain,sunshine --critical=8.3
1770
1771       For MRTG output, reports a 1 or a 0 indicating success or failure on
1772       the first line. The fourth line indicates the current version. The
1773       version must be provided via the "--mrtg" option.
1774
1775   wal_files
1776       ("symlink: check_postgres_wal_files") Checks how many WAL files exist
1777       in the pg_xlog directory (PostgreSQL 10 and later" pg_wal), which is
1778       found off of your data_directory, sometimes as a symlink to another
1779       physical disk for performance reasons. If the --lsfunc option is not
1780       used then this action must be run as a superuser, in order to access
1781       the contents of the pg_xlog directory. The minimum version to use this
1782       action is Postgres 8.1. The --warning and --critical options are simply
1783       the number of files in the pg_xlog directory. What number to set this
1784       to will vary, but a general guideline is to put a number slightly
1785       higher than what is normally there, to catch problems early.
1786
1787       Normally, WAL files are closed and then re-used, but a long-running
1788       open transaction, or a faulty archive_command script, may cause
1789       Postgres to create too many files. Ultimately, this will cause the disk
1790       they are on to run out of space, at which point Postgres will shut
1791       down.
1792
1793       To avoid connecting as a database superuser, a wrapper function around
1794       "pg_ls_dir()" should be defined as a superuser with SECURITY DEFINER,
1795       and the --lsfunc option used. This example function, if defined by a
1796       superuser, will allow the script to connect as a normal user nagios
1797       with --lsfunc=ls_xlog_dir
1798
1799         BEGIN;
1800         CREATE FUNCTION ls_xlog_dir()
1801             RETURNS SETOF TEXT
1802             AS $$ SELECT pg_ls_dir('pg_xlog') $$
1803             LANGUAGE SQL
1804             SECURITY DEFINER;
1805         REVOKE ALL ON FUNCTION ls_xlog_dir() FROM PUBLIC;
1806         GRANT EXECUTE ON FUNCTION ls_xlog_dir() to nagios;
1807         COMMIT;
1808
1809       Example 1: Check that the number of ready WAL files is 10 or less on
1810       host "pluto", using a wrapper function "ls_xlog_dir" to avoid the need
1811       for superuser permissions
1812
1813         check_postgres_archive_ready --host=pluto --critical=10 --lsfunc=ls_xlog_dir
1814
1815       For MRTG output, reports the number of WAL files on line 1.
1816
1817   rebuild_symlinks
1818   rebuild_symlinks_force
1819       This action requires no other arguments, and does not connect to any
1820       databases, but simply creates symlinks in the current directory for
1821       each action, in the form check_postgres_<action_name>.  If the file
1822       already exists, it will not be overwritten. If the action is
1823       rebuild_symlinks_force, then symlinks will be overwritten. The option
1824       --symlinks is a shorter way of saying --action=rebuild_symlinks
1825

BASIC FILTERING

1827       The options --include and --exclude can be combined to limit which
1828       things are checked, depending on the action. The name of the database
1829       can be filtered when using the following actions: backends,
1830       database_size, locks, query_time, txn_idle, and txn_time.  The name of
1831       a relation can be filtered when using the following actions: bloat,
1832       index_size, table_size, relation_size, last_vacuum, last_autovacuum,
1833       last_analyze, and last_autoanalyze.  The name of a setting can be
1834       filtered when using the settings_checksum action.  The name of a file
1835       system can be filtered when using the disk_space action.
1836
1837       If only an include option is given, then ONLY those entries that match
1838       will be checked. However, if given both exclude and include, the
1839       exclusion is done first, and the inclusion after, to reinstate things
1840       that may have been excluded. Both --include and --exclude can be given
1841       multiple times, and/or as comma-separated lists. A leading tilde will
1842       match the following word as a regular expression.
1843
1844       To match a schema, end the search term with a single period. Leading
1845       tildes can be used for schemas as well.
1846
1847       Be careful when using filtering: an inclusion rule on the backends, for
1848       example, may report no problems not only because the matching database
1849       had no backends, but because you misspelled the name of the database!
1850
1851       Examples:
1852
1853       Only checks items named pg_class:
1854
1855        --include=pg_class
1856
1857       Only checks items containing the letters 'pg_':
1858
1859        --include=~pg_
1860
1861       Only check items beginning with 'pg_':
1862
1863        --include=~^pg_
1864
1865       Exclude the item named 'test':
1866
1867        --exclude=test
1868
1869       Exclude all items containing the letters 'test:
1870
1871        --exclude=~test
1872
1873       Exclude all items in the schema 'pg_catalog':
1874
1875        --exclude='pg_catalog.'
1876
1877       Exclude all items containing the letters 'ace', but allow the item
1878       'faceoff':
1879
1880        --exclude=~ace --include=faceoff
1881
1882       Exclude all items which start with the letters 'pg_', which contain the
1883       letters 'slon', or which are named 'sql_settings' or 'green'.
1884       Specifically check items with the letters 'prod' in their names, and
1885       always check the item named 'pg_relname':
1886
1887        --exclude=~^pg_,~slon,sql_settings --exclude=green --include=~prod,pg_relname
1888

USER NAME FILTERING

1890       The options --includeuser and --excludeuser can be used on some actions
1891       to only examine database objects owned by (or not owned by) one or more
1892       users.  An --includeuser option always trumps an --excludeuser option.
1893       You can give each option more than once for multiple users, or you can
1894       give a comma-separated list. The actions that currently use these
1895       options are:
1896
1897       database_size
1898       last_analyze
1899       last_autoanalyze
1900       last_vacuum
1901       last_autovacuum
1902       query_time
1903       relation_size
1904       txn_time
1905
1906       Examples:
1907
1908       Only check items owned by the user named greg:
1909
1910        --includeuser=greg
1911
1912       Only check items owned by either watson or crick:
1913
1914        --includeuser=watson,crick
1915
1916       Only check items owned by crick,franklin, watson, or wilkins:
1917
1918        --includeuser=watson --includeuser=franklin --includeuser=crick,wilkins
1919
1920       Check all items except for those belonging to the user scott:
1921
1922        --excludeuser=scott
1923

TEST MODE

1925       To help in setting things up, this program can be run in a "test mode"
1926       by specifying the --test option. This will perform some basic tests to
1927       make sure that the databases can be contacted, and that certain per-
1928       action prerequisites are met, such as whether the user is a superuser,
1929       if the version of Postgres is new enough, and if stats_row_level is
1930       enabled.
1931

FILES

1933       In addition to command-line configurations, you can put any options
1934       inside of a file. The file .check_postgresrc in the current directory
1935       will be used if found. If not found, then the file ~/.check_postgresrc
1936       will be used. Finally, the file /etc/check_postgresrc will be used if
1937       available.  The format of the file is option = value, one per line. Any
1938       line starting with a '#' will be skipped.  Any values loaded from a
1939       check_postgresrc file will be overwritten by command-line options. All
1940       check_postgresrc files can be ignored by supplying a
1941       "--no-checkpostgresrc" argument.
1942

ENVIRONMENT VARIABLES

1944       The environment variable $ENV{HOME} is used to look for a
1945       .check_postgresrc file.  The environment variable $ENV{PGBINDIR} is
1946       used to look for PostgreSQL binaries.
1947

TIPS AND TRICKS

1949       Since this program uses the psql program, make sure it is accessible to
1950       the user running the script. If run as a cronjob, this often means
1951       modifying the PATH environment variable.
1952
1953       If you are using Nagios in embedded Perl mode, use the "--action"
1954       argument instead of symlinks, so that the plugin only gets compiled one
1955       time.
1956

DEPENDENCIES

1958       Access to a working version of psql, and the following very standard
1959       Perl modules:
1960
1961       Cwd
1962       Getopt::Long
1963       File::Basename
1964       File::Temp
1965       Time::HiRes (if $opt{showtime} is set to true, which is the default)
1966
1967       The "settings_checksum" action requires the Digest::MD5 module.
1968
1969       The "checkpoint" action requires the Date::Parse module.
1970
1971       Some actions require access to external programs. If psql is not
1972       explicitly specified, the command "which" is used to find it. The
1973       program "/bin/df" is needed by the "disk_space" action.
1974

DEVELOPMENT

1976       Development happens using the git system. You can clone the latest
1977       version by doing:
1978
1979        https://github.com/bucardo/check_postgres
1980        git clone https://github.com/bucardo/check_postgres.git
1981

MAILING LIST

1983       Three mailing lists are available. For discussions about the program,
1984       bug reports, feature requests, and commit notices, send email to
1985       check_postgres@bucardo.org
1986
1987       https://mail.endcrypt.com/mailman/listinfo/check_postgres
1988
1989       A low-volume list for announcement of new versions and important
1990       notices is the 'check_postgres-announce' list:
1991
1992       https://mail.endcrypt.com/mailman/listinfo/check_postgres-announce
1993
1994       Source code changes (via git-commit) are sent to the
1995       'check_postgres-commit' list:
1996
1997       https://mail.endcrypt.com/mailman/listinfo/check_postgres-commit
1998

HISTORY

2000       Items not specifically attributed are by GSM (Greg Sabino Mullane).
2001
2002       Version 2.24.0 Released May 30, 2018
2003             Support new_version_pg for PG10
2004               (Michael Pirogov)
2005
2006             Option to skip CYCLE sequences in action sequence
2007               (Christoph Moench-Tegeder)
2008
2009             Output per-database perfdata for pgbouncer pool checks
2010               (George Hansper)
2011
2012             German message translations
2013               (Holger Jacobs)
2014
2015             Consider only client backends in query_time and friends
2016               (David Christensen)
2017
2018       Version 2.23.0 Released October 31, 2017
2019             Support PostgreSQL 10.
2020               (David Christensen, Christoph Berg)
2021
2022             Change table_size to use pg_table_size() on 9.0+, i.e. include the TOAST
2023             table size in the numbers reported. Add new actions indexes_size and
2024             total_relation_size, using the respective pg_indexes_size() and
2025             pg_total_relation_size() functions. All size checks will now also check
2026             materialized views where applicable.
2027               (Christoph Berg)
2028
2029             Connection errors are now always critical, not unknown.
2030               (Christoph Berg)
2031
2032             New action replication_slots checking if logical or physical replication
2033             slots have accumulated too much data
2034               (Glyn Astill)
2035
2036             Multiple same_schema improvements
2037               (Glyn Astill)
2038
2039             Add Spanish message translations
2040               (Luis Vazquez)
2041
2042             Allow a wrapper function to run wal_files and archive_ready actions as
2043             non-superuser
2044               (Joshua Elsasser)
2045
2046             Add some defensive casting to the bloat query
2047               (Greg Sabino Mullane)
2048
2049             Invoke psql with option -X
2050               (Peter Eisentraut)
2051
2052             Update postgresql.org URLs to use https.
2053               (Magnus Hagander)
2054
2055             check_txn_idle: Don't fail when query contains 'disabled' word
2056               (Marco Nenciarini)
2057
2058             check_txn_idle: Use state_change instead of query_start.
2059               (Sebastian Webber)
2060
2061             check_hot_standby_delay: Correct extra space in perfdata
2062               (Adrien Nayrat)
2063
2064             Remove \r from psql output as it can confuse some regexes
2065               (Greg Sabino Mullane)
2066
2067             Sort failed jobs in check_pgagent_jobs for stable output.
2068               (Christoph Berg)
2069
2070       Version 2.22.0 June 30, 2015
2071             Add xact timestamp support to hot_standby_delay.
2072             Allow the hot_standby_delay check to accept xlog byte position or
2073             timestamp lag intervals as thresholds, or even both at the same time.
2074               (Josh Williams)
2075
2076             Query all sequences per DB in parallel for action=sequence.
2077               (Christoph Berg)
2078
2079             Fix bloat check to use correct SQL depending on the server version.
2080               (Adrian Vondendriesch)
2081
2082             Show actual long-running query in query_time output
2083               (Peter Eisentraut)
2084
2085             Add explicit ORDER BY to the slony_status check to get the most lagged server.
2086               (Jeff Frost)
2087
2088             Improved multi-slave support in replicate_row.
2089               (Andrew Yochum)
2090
2091             Change the way tables are quoted in replicate_row.
2092               (Glyn Astill)
2093
2094             Don't swallow space before the -c flag when reporting errors
2095               (Jeff Janes)
2096
2097             Fix and extend hot_standby_delay documentation
2098               (Michael Renner)
2099
2100             Declare POD encoding to be utf8.
2101               (Christoph Berg)
2102
2103       Version 2.21.0 September 24, 2013
2104             Fix issue with SQL steps in check_pgagent_jobs for sql steps which perform deletes
2105               (Rob Emery via github pull)
2106
2107             Install man page in section 1.
2108               (Peter Eisentraut, bug 53, github issue 26)
2109
2110             Order lock types in check_locks output to make the ordering predictable;
2111             setting SKIP_NETWORK_TESTS will skip the new_version tests; other minor test
2112             suite fixes.
2113               (Christoph Berg)
2114
2115             Fix same_schema check on 9.3 by ignoring relminmxid differences in pg_class
2116               (Christoph Berg)
2117
2118       Version 2.20.1 June 24, 2013
2119             Make connection check failures return CRITICAL not UNKNOWN
2120               (Dominic Hargreaves)
2121
2122             Fix --reverse option when using string comparisons in custom queries
2123               (Nathaniel Waisbrot)
2124
2125             Compute correct 'totalwastedbytes' in the bloat query
2126               (Michael Renner)
2127
2128             Do not use pg_stats "inherited" column in bloat query, if the
2129               database is 8.4 or older. (Greg Sabino Mullane, per bug 121)
2130
2131             Remove host reordering in hot_standby_delay check
2132               (Josh Williams, with help from Jacobo Blasco)
2133
2134             Better output for the "simple" flag
2135               (Greg Sabino Mullane)
2136
2137             Force same_schema to ignore the 'relallvisible' column
2138               (Greg Sabino Mullane)
2139
2140       Version 2.20.0 March 13, 2013
2141             Add check for pgagent jobs (David E. Wheeler)
2142
2143             Force STDOUT to use utf8 for proper output
2144               (Greg Sabino Mullane; reported by Emmanuel Lesouef)
2145
2146             Fixes for Postgres 9.2: new pg_stat_activity view,
2147               and use pg_tablespace_location, (Josh Williams)
2148
2149             Allow for spaces in item lists when doing same_schema.
2150
2151             Allow txn_idle to work again for < 8.3 servers by switching to query_time.
2152
2153             Fix the check_bloat SQL to take inherited tables into account,
2154               and assume 2k for non-analyzed columns. (Geert Pante)
2155
2156             Cache sequence information to speed up same_schema runs.
2157
2158             Fix --excludeuser in check_txn_idle (Mika Eloranta)
2159
2160             Fix user clause handling in check_txn_idle (Michael van Bracht)
2161
2162             Adjust docs to show colon as a better separator inside args for locks
2163               (Charles Sprickman)
2164
2165             Fix undefined $SQL2 error in check_txn_idle [github issue 16] (Patric Bechtel)
2166
2167             Prevent "uninitialized value" warnings when showing the port (Henrik Ahlgren)
2168
2169             Do not assume everyone has a HOME [github issue 23]
2170
2171       Version 2.19.0 January 17, 2012
2172             Add the --assume-prod option (Cédric Villemain)
2173
2174             Add the cluster_id check (Cédric Villemain)
2175
2176             Improve settings_checksum and checkpoint tests (Cédric Villemain)
2177
2178             Do not do an inner join to pg_user when checking database size
2179               (Greg Sabino Mullane; reported by Emmanuel Lesouef)
2180
2181             Use the full path when getting sequence information for same_schema.
2182               (Greg Sabino Mullane; reported by Cindy Wise)
2183
2184             Fix the formula for calculating xlog positions (Euler Taveira de Oliveira)
2185
2186             Better ordering of output for bloat check - make indexes as important
2187               as tables (Greg Sabino Mullane; reported by Jens Wilke)
2188
2189             Show the dbservice if it was used at top of same_schema output
2190               (Mike Blackwell)
2191
2192             Better installation paths (Greg Sabino Mullane, per bug 53)
2193
2194       Version 2.18.0 October 2, 2011
2195             Redo the same_schema action. Use new --filter argument for all filtering.
2196             Allow comparisons between any number of databases.
2197             Remove the dbname2, dbport2, etc. arguments.
2198             Allow comparison of the same db over time.
2199
2200             Swap db1 and db2 if the slave is 1 for the hot standby check (David E. Wheeler)
2201
2202             Allow multiple --schema arguments for the slony_status action (GSM and Jehan-Guillaume de Rorthais)
2203
2204             Fix ORDER BY in the last vacuum/analyze action (Nicolas Thauvin)
2205
2206             Fix check_hot_standby_delay perfdata output (Nicolas Thauvin)
2207
2208             Look in the correct place for the .ready files with the archive_ready action (Nicolas Thauvin)
2209
2210             New action: commitratio (Guillaume Lelarge)
2211
2212             New action: hitratio (Guillaume Lelarge)
2213
2214             Make sure --action overrides the symlink naming trick.
2215
2216             Set defaults for archive_ready and wal_files (Thomas Guettler, GSM)
2217
2218             Better output for wal_files and archive_ready (GSM)
2219
2220             Fix warning when client_port set to empty string (bug #79)
2221
2222             Account for "empty row" in -x output (i.e. source of functions).
2223
2224             Fix some incorrectly named data fields (Andy Lester)
2225
2226             Expand the number of pgbouncer actions (Ruslan Kabalin)
2227
2228             Give detailed information and refactor txn_idle, txn_time, and query_time
2229               (Per request from bug #61)
2230
2231             Set maxalign to 8 in the bloat check if box identified as '64-bit'
2232               (Michel Sijmons, bug #66)
2233
2234             Support non-standard version strings in the bloat check.
2235               (Michel Sijmons and Gurjeet Singh, bug #66)
2236
2237             Do not show excluded databases in some output (Ruslan Kabalin)
2238
2239             Allow "and", "or" inside arguments (David E. Wheeler)
2240
2241             Add the "new_version_box" action.
2242
2243             Fix psql version regex (Peter Eisentraut, bug #69)
2244
2245             Add the --assume-standby-mode option (Ruslan Kabalin)
2246
2247             Note that txn_idle and query_time require 8.3 (Thomas Guettler)
2248
2249             Standardize and clean up all perfdata output (bug #52)
2250
2251             Exclude "idle in transaction" from the query_time check (bug #43)
2252
2253             Fix the perflimit for the bloat action (bug #50)
2254
2255             Clean up the custom_query action a bit.
2256
2257             Fix space in perfdata for hot_standby_delay action (Nicolas Thauvin)
2258
2259             Handle undef percents in check_fsm_relations (Andy Lester)
2260
2261             Fix typo in dbstats action (Stas Vitkovsky)
2262
2263             Fix MRTG for last vacuum and last_analyze actions.
2264
2265       Version 2.17.0 no public release
2266       Version 2.16.0 January 20, 2011
2267             Add new action 'hot_standby_delay' (Nicolas Thauvin)
2268             Add cache-busting for the version-grabbing utilities.
2269             Fix problem with going to next method for new_version_pg
2270               (Greg Sabino Mullane, reported by Hywel Mallett in bug #65)
2271             Allow /usr/local/etc as an alternative location for the
2272               check_postgresrc file (Hywel Mallett)
2273             Do not use tgisconstraint in same_schema if Postgres >= 9
2274               (Guillaume Lelarge)
2275
2276       Version 2.15.4 January 3, 2011
2277             Fix warning when using symlinks
2278               (Greg Sabino Mullane, reported by Peter Eisentraut in bug #63)
2279
2280       Version 2.15.3 December 30, 2010
2281             Show OK for no matching txn_idle entries.
2282
2283       Version 2.15.2 December 28, 2010
2284             Better formatting of sizes in the bloat action output.
2285
2286             Remove duplicate perfs in bloat action output.
2287
2288       Version 2.15.1 December 27, 2010
2289             Fix problem when examining items in pg_settings (Greg Sabino Mullane)
2290
2291             For connection test, return critical, not unknown, on FATAL errors
2292               (Greg Sabino Mullane, reported by Peter Eisentraut in bug #62)
2293
2294       Version 2.15.0 November 8, 2010
2295             Add --quiet argument to suppress output on OK Nagios results
2296             Add index comparison for same_schema (Norman Yamada and Greg Sabino Mullane)
2297             Use $ENV{PGSERVICE} instead of "service=" to prevent problems (Guillaume Lelarge)
2298             Add --man option to show the entire manual. (Andy Lester)
2299             Redo the internal run_command() sub to use -x and hashes instead of regexes.
2300             Fix error in custom logic (Andreas Mager)
2301             Add the "pgbouncer_checksum" action (Guillaume Lelarge)
2302             Fix regex to work on WIN32 for check_fsm_relations and check_fsm_pages (Luke Koops)
2303             Don't apply a LIMIT when using --exclude on the bloat action (Marti Raudsepp)
2304             Change the output of query_time to show pid,user,port, and address (Giles Westwood)
2305             Fix to show database properly when using slony_status (Guillaume Lelarge)
2306             Allow warning items for same_schema to be comma-separated (Guillaume Lelarge)
2307             Constraint definitions across Postgres versions match better in same_schema.
2308             Work against "EnterpriseDB" databases (Sivakumar Krishnamurthy and Greg Sabino Mullane)
2309             Separate perfdata with spaces (Jehan-Guillaume (ioguix) de Rorthais)
2310             Add new action "archive_ready" (Jehan-Guillaume (ioguix) de Rorthais)
2311
2312       Version 2.14.3 (March 1, 2010)
2313             Allow slony_status action to handle more than one slave.
2314             Use commas to separate function args in same_schema output (Robert Treat)
2315
2316       Version 2.14.2 (February 18, 2010)
2317             Change autovac_freeze default warn/critical back to 90%/95% (Robert Treat)
2318             Put all items one-per-line for relation size actions if --verbose=1
2319
2320       Version 2.14.1 (February 17, 2010)
2321             Don't use $^T in logfile check, as script may be long-running
2322             Change the error string for the logfile action for easier exclusion
2323               by programs like tail_n_mail
2324
2325       Version 2.14.0 (February 11, 2010)
2326             Added the 'slony_status' action.
2327             Changed the logfile sleep from 0.5 to 1, as 0.5 gets rounded to 0 on some boxes!
2328
2329       Version 2.13.2 (February 4, 2010)
2330             Allow timeout option to be used for logtime 'sleep' time.
2331
2332       Version 2.13.2 (February 4, 2010)
2333             Show offending database for query_time action.
2334             Apply perflimit to main output for sequence action.
2335             Add 'noowner' option to same_schema action.
2336             Raise sleep timeout for logfile check to 15 seconds.
2337
2338       Version 2.13.1 (February 2, 2010)
2339             Fix bug preventing column constraint differences from 2 > 1 for same_schema from being shown.
2340             Allow aliases 'dbname1', 'dbhost1', 'dbport1',etc.
2341             Added "nolanguage" as a filter for the same_schema option.
2342             Don't track "generic" table constraints (e.. $1, $2) using same_schema
2343
2344       Version 2.13.0 (January 29, 2010)
2345             Allow "nofunctions" as a filter for the same_schema option.
2346             Added "noperm" as a filter for the same_schema option.
2347             Ignore dropped columns when considered positions for same_schema (Guillaume Lelarge)
2348
2349       Version 2.12.1 (December 3, 2009)
2350             Change autovac_freeze default warn/critical from 90%/95% to 105%/120% (Marti Raudsepp)
2351
2352       Version 2.12.0 (December 3, 2009)
2353             Allow the temporary directory to be specified via the "tempdir" argument,
2354               for systems that need it (e.g. /tmp is not owned by root).
2355             Fix so old versions of Postgres (< 8.0) use the correct default database (Giles Westwood)
2356             For "same_schema" trigger mismatches, show the attached table.
2357             Add the new_version_bc check for Bucardo version checking.
2358             Add database name to perf output for last_vacuum|analyze (Guillaume Lelarge)
2359             Fix for bloat action against old versions of Postgres without the 'block_size' param.
2360
2361       Version 2.11.1 (August 27, 2009)
2362             Proper Nagios output for last_vacuum|analyze actions. (Cédric Villemain)
2363             Proper Nagios output for locks action. (Cédric Villemain)
2364             Proper Nagios output for txn_wraparound action. (Cédric Villemain)
2365             Fix for constraints with embedded newlines for same_schema.
2366             Allow --exclude for all items when using same_schema.
2367
2368       Version 2.11.0 (August 23, 2009)
2369             Add Nagios perf output to the wal_files check (Cédric Villemain)
2370             Add support for .check_postgresrc, per request from Albe Laurenz.
2371             Allow list of web fetch methods to be changed with the --get_method option.
2372             Add support for the --language argument, which overrides any ENV.
2373             Add the --no-check_postgresrc flag.
2374             Ensure check_postgresrc options are completely overridden by command-line options.
2375             Fix incorrect warning > critical logic in replicate_rows (Glyn Astill)
2376
2377       Version 2.10.0 (August 3, 2009)
2378             For same_schema, compare view definitions, and compare languages.
2379             Make script into a global executable via the Makefile.PL file.
2380             Better output when comparing two databases.
2381             Proper Nagios output syntax for autovac_freeze and backends checks (Cédric Villemain)
2382
2383       Version 2.9.5 (July 24, 2009)
2384             Don't use a LIMIT in check_bloat if --include is used. Per complaint from Jeff Frost.
2385
2386       Version 2.9.4 (July 21, 2009)
2387             More French translations (Guillaume Lelarge)
2388
2389       Version 2.9.3 (July 14, 2009)
2390             Quote dbname in perf output for the backends check. (Davide Abrigo)
2391             Add 'fetch' as an alternative method for new_version checks, as this
2392               comes by default with FreeBSD. (Hywel Mallett)
2393
2394       Version 2.9.2 (July 12, 2009)
2395             Allow dots and dashes in database name for the backends check (Davide Abrigo)
2396             Check and display the database for each match in the bloat check (Cédric Villemain)
2397             Handle 'too many connections' FATAL error in the backends check with a critical,
2398               rather than a generic error (Greg, idea by Jürgen Schulz-Brüssel)
2399             Do not allow perflimit to interfere with exclusion rules in the vacuum and
2400               analyze tests. (Greg, bug reported by Jeff Frost)
2401
2402       Version 2.9.1 (June 12, 2009)
2403             Fix for multiple databases with the check_bloat action (Mark Kirkwood)
2404             Fixes and improvements to the same_schema action (Jeff Boes)
2405             Write tests for same_schema, other minor test fixes (Jeff Boes)
2406
2407       Version 2.9.0 (May 28, 2009)
2408             Added the same_schema action (Greg)
2409
2410       Version 2.8.1 (May 15, 2009)
2411             Added timeout via statement_timeout in addition to perl alarm (Greg)
2412
2413       Version 2.8.0 (May 4, 2009)
2414             Added internationalization support (Greg)
2415             Added the 'disabled_triggers' check (Greg)
2416             Added the 'prepared_txns' check (Greg)
2417             Added the 'new_version_cp' and 'new_version_pg' checks (Greg)
2418             French translations (Guillaume Lelarge)
2419             Make the backends search return ok if no matches due to inclusion rules,
2420               per report by Guillaume Lelarge (Greg)
2421             Added comprehensive unit tests (Greg, Jeff Boes, Selena Deckelmann)
2422             Make fsm_pages and fsm_relations handle 8.4 servers smoothly. (Greg)
2423             Fix missing 'upd' field in show_dbstats (Andras Fabian)
2424             Allow ENV{PGCONTROLDATA} and ENV{PGBINDIR}. (Greg)
2425             Add various Perl module infrastructure (e.g. Makefile.PL) (Greg)
2426             Fix incorrect regex in txn_wraparound (Greg)
2427             For txn_wraparound: consistent ordering and fix duplicates in perf output (Andras Fabian)
2428             Add in missing exabyte regex check (Selena Deckelmann)
2429             Set stats to zero if we bail early due to USERWHERECLAUSE (Andras Fabian)
2430             Add additional items to dbstats output (Andras Fabian)
2431             Remove --schema option from the fsm_ checks. (Greg Mullane and Robert Treat)
2432             Handle case when ENV{PGUSER} is set. (Andy Lester)
2433             Many various fixes. (Jeff Boes)
2434             Fix --dbservice: check version and use ENV{PGSERVICE} for old versions (Cédric Villemain)
2435
2436       Version 2.7.3 (February 10, 2009)
2437             Make the sequence action check if sequence being used for a int4 column and
2438             react appropriately. (Michael Glaesemann)
2439
2440       Version 2.7.2 (February 9, 2009)
2441             Fix to prevent multiple groupings if db arguments given.
2442
2443       Version 2.7.1 (February 6, 2009)
2444             Allow the -p argument for port to work again.
2445
2446       Version 2.7.0 (February 4, 2009)
2447             Do not require a connection argument, but use defaults and ENV variables when
2448               possible: PGHOST, PGPORT, PGUSER, PGDATABASE.
2449
2450       Version 2.6.1 (February 4, 2009)
2451             Only require Date::Parse to be loaded if using the checkpoint action.
2452
2453       Version 2.6.0 (January 26, 2009)
2454             Add the 'checkpoint' action.
2455
2456       Version 2.5.4 (January 7, 2009)
2457             Better checking of $opt{dbservice} structure (Cédric Villemain)
2458             Fix time display in timesync action output (Selena Deckelmann)
2459             Fix documentation typos (Josh Tolley)
2460
2461       Version 2.5.3 (December 17, 2008)
2462             Minor fix to regex in verify_version (Lee Jensen)
2463
2464       Version 2.5.2 (December 16, 2008)
2465             Minor documentation tweak.
2466
2467       Version 2.5.1 (December 11, 2008)
2468             Add support for --noidle flag to prevent backends action from counting idle processes.
2469             Patch by Selena Deckelmann.
2470
2471             Fix small undefined warning when not using --dbservice.
2472
2473       Version 2.5.0 (December 4, 2008)
2474             Add support for the pg_Service.conf file with the --dbservice option.
2475
2476       Version 2.4.3 (November 7, 2008)
2477             Fix options for replicate_row action, per report from Jason Gordon.
2478
2479       Version 2.4.2 (November 6, 2008)
2480             Wrap File::Temp::cleanup() calls in eval, in case File::Temp is an older version.
2481             Patch by Chris Butler.
2482
2483       Version 2.4.1 (November 5, 2008)
2484             Cast numbers to numeric to support sequences ranges > bigint in check_sequence action.
2485             Thanks to Scott Marlowe for reporting this.
2486
2487       Version 2.4.0 (October 26, 2008)
2488            Add Cacti support with the dbstats action.
2489            Pretty up the time output for last vacuum and analyze actions.
2490            Show the percentage of backends on the check_backends action.
2491
2492       Version 2.3.10 (October 23, 2008)
2493            Fix minor warning in action check_bloat with multiple databases.
2494            Allow warning to be greater than critical when using the --reverse option.
2495            Support the --perflimit option for the check_sequence action.
2496
2497       Version 2.3.9 (October 23, 2008)
2498            Minor tweak to way we store the default port.
2499
2500       Version 2.3.8 (October 21, 2008)
2501            Allow the default port to be changed easily.
2502            Allow transform of simple output by MB, GB, etc.
2503
2504       Version 2.3.7 (October 14, 2008)
2505            Allow multiple databases in 'sequence' action. Reported by Christoph Zwerschke.
2506
2507       Version 2.3.6  (October 13, 2008)
2508            Add missing $schema to check_fsm_pages. (Robert Treat)
2509
2510       Version 2.3.5 (October 9, 2008)
2511            Change option 'checktype' to 'valtype' to prevent collisions with -c[ritical]
2512            Better handling of errors.
2513
2514       Version 2.3.4 (October 9, 2008)
2515            Do explicit cleanups of the temp directory, per problems reported by sb@nnx.com.
2516
2517       Version 2.3.3 (October 8, 2008)
2518            Account for cases where some rounding queries give -0 instead of 0.
2519            Thanks to Glyn Astill for helping to track this down.
2520
2521       Version 2.3.2 (October 8, 2008)
2522            Always quote identifiers in check_replicate_row action.
2523
2524       Version 2.3.1 (October 7, 2008)
2525            Give a better error if one of the databases cannot be reached.
2526
2527       Version 2.3.0 (October 4, 2008)
2528            Add the "sequence" action, thanks to Gavin M. Roy for the idea.
2529            Fix minor problem with autovac_freeze action when using MRTG output.
2530            Allow output argument to be case-insensitive.
2531            Documentation fixes.
2532
2533       Version 2.2.4 (October 3, 2008)
2534            Fix some minor typos
2535
2536       Version 2.2.3 (October 1, 2008)
2537            Expand range of allowed names for --repinfo argument (Glyn Astill)
2538            Documentation tweaks.
2539
2540       Version 2.2.2 (September 30, 2008)
2541            Fixes for minor output and scoping problems.
2542
2543       Version 2.2.1 (September 28, 2008)
2544            Add MRTG output to fsm_pages and fsm_relations.
2545            Force error messages to one-line for proper Nagios output.
2546            Check for invalid prereqs on failed command. From conversations with Euler Taveira de Oliveira.
2547            Tweak the fsm_pages formula a little.
2548
2549       Version 2.2.0 (September 25, 2008)
2550            Add fsm_pages and fsm_relations actions. (Robert Treat)
2551
2552       Version 2.1.4 (September 22, 2008)
2553            Fix for race condition in txn_time action.
2554            Add --debugoutput option.
2555
2556       Version 2.1.3 (September 22, 2008)
2557            Allow alternate arguments "dbhost" for "host" and "dbport" for "port".
2558            Output a zero as default value for second line of MRTG output.
2559
2560       Version 2.1.2 (July 28, 2008)
2561            Fix sorting error in the "disk_space" action for non-Nagios output.
2562            Allow --simple as a shortcut for --output=simple.
2563
2564       Version 2.1.1 (July 22, 2008)
2565            Don't check databases with datallowconn false for the "autovac_freeze" action.
2566
2567       Version 2.1.0 (July 18, 2008)
2568            Add the "autovac_freeze" action, thanks to Robert Treat for the idea and design.
2569            Put an ORDER BY on the "txn_wraparound" action.
2570
2571       Version 2.0.1 (July 16, 2008)
2572            Optimizations to speed up the "bloat" action quite a bit.
2573            Fix "version" action to not always output in mrtg mode.
2574
2575       Version 2.0.0 (July 15, 2008)
2576            Add support for MRTG and "simple" output options.
2577            Many small improvements to nearly all actions.
2578
2579       Version 1.9.1 (June 24, 2008)
2580            Fix an error in the bloat SQL in 1.9.0
2581            Allow percentage arguments to be over 99%
2582            Allow percentages in the bloat --warning and --critical (thanks to Robert Treat for the idea)
2583
2584       Version 1.9.0 (June 22, 2008)
2585            Don't include information_schema in certain checks. (Jeff Frost)
2586            Allow --include and --exclude to use schemas by using a trailing period.
2587
2588       Version 1.8.5 (June 22, 2008)
2589            Output schema name before table name where appropriate.
2590            Thanks to Jeff Frost.
2591
2592       Version 1.8.4 (June 19, 2008)
2593            Better detection of problems in --replicate_row.
2594
2595       Version 1.8.3 (June 18, 2008)
2596            Fix 'backends' action: there may be no rows in pg_stat_activity, so run a second
2597              query if needed to find the max_connections setting.
2598            Thanks to Jeff Frost for the bug report.
2599
2600       Version 1.8.2 (June 10, 2008)
2601            Changes to allow working under Nagios' embedded Perl mode. (Ioannis Tambouras)
2602
2603       Version 1.8.1 (June 9, 2008)
2604            Allow 'bloat' action to work on Postgres version 8.0.
2605            Allow for different commands to be run for each action depending on the server version.
2606            Give better warnings when running actions not available on older Postgres servers.
2607
2608       Version 1.8.0 (June 3, 2008)
2609            Add the --reverse option to the custom_query action.
2610
2611       Version 1.7.1 (June 2, 2008)
2612            Fix 'query_time' action: account for race condition in which zero rows appear in pg_stat_activity.
2613            Thanks to Dustin Black for the bug report.
2614
2615       Version 1.7.0 (May 11, 2008)
2616            Add --replicate_row action
2617
2618       Version 1.6.1 (May 11, 2008)
2619            Add --symlinks option as a shortcut to --action=rebuild_symlinks
2620
2621       Version 1.6.0 (May 11, 2008)
2622            Add the custom_query action.
2623
2624       Version 1.5.2 (May 2, 2008)
2625            Fix problem with too eager creation of custom pgpass file.
2626
2627       Version 1.5.1 (April 17, 2008)
2628            Add example Nagios configuration settings (Brian A. Seklecki)
2629
2630       Version 1.5.0 (April 16, 2008)
2631            Add the --includeuser and --excludeuser options. Documentation cleanup.
2632
2633       Version 1.4.3 (April 16, 2008)
2634            Add in the 'output' concept for future support of non-Nagios programs.
2635
2636       Version 1.4.2 (April 8, 2008)
2637            Fix bug preventing --dbpass argument from working (Robert Treat).
2638
2639       Version 1.4.1 (April 4, 2008)
2640            Minor documentation fixes.
2641
2642       Version 1.4.0 (April 2, 2008)
2643            Have 'wal_files' action use pg_ls_dir (idea by Robert Treat).
2644            For last_vacuum and last_analyze, respect autovacuum effects, add separate
2645              autovacuum checks (ideas by Robert Treat).
2646
2647       Version 1.3.1 (April 2, 2008)
2648            Have txn_idle use query_start, not xact_start.
2649
2650       Version 1.3.0 (March 23, 2008)
2651            Add in txn_idle and txn_time actions.
2652
2653       Version 1.2.0 (February 21, 2008)
2654            Add the 'wal_files' action, which counts the number of WAL files
2655              in your pg_xlog directory.
2656            Fix some typos in the docs.
2657            Explicitly allow -v as an argument.
2658            Allow for a null syslog_facility in the 'logfile' action.
2659
2660       Version 1.1.2 (February 5, 2008)
2661            Fix error preventing --action=rebuild_symlinks from working.
2662
2663       Version 1.1.1 (February 3, 2008)
2664            Switch vacuum and analyze date output to use 'DD', not 'D'. (Glyn Astill)
2665
2666       Version 1.1.0 (December 16, 2008)
2667            Fixes, enhancements, and performance tracking.
2668            Add performance data tracking via --showperf and --perflimit
2669            Lots of refactoring and cleanup of how actions handle arguments.
2670            Do basic checks to figure out syslog file for 'logfile' action.
2671            Allow for exact matching of beta versions with 'version' action.
2672            Redo the default arguments to only populate when neither 'warning' nor 'critical' is provided.
2673            Allow just warning OR critical to be given for the 'timesync' action.
2674            Remove 'redirect_stderr' requirement from 'logfile' due to 8.3 changes.
2675            Actions 'last_vacuum' and 'last_analyze' are 8.2 only (Robert Treat)
2676
2677       Version 1.0.16 (December 7, 2007)
2678            First public release, December 2007
2679

BUGS AND LIMITATIONS

2681       The index bloat size optimization is rough.
2682
2683       Some actions may not work on older versions of Postgres (before 8.0).
2684
2685       Please report any problems to check_postgres@bucardo.org
2686

AUTHOR

2688       Greg Sabino Mullane <greg@endpoint.com>
2689

NAGIOS EXAMPLES

2691       Some example Nagios configuration settings using this script:
2692
2693        define command {
2694            command_name    check_postgres_size
2695            command_line    $USER2$/check_postgres.pl -H $HOSTADDRESS$ -u pgsql -db postgres --action database_size -w $ARG1$ -c $ARG2$
2696        }
2697
2698        define command {
2699            command_name    check_postgres_locks
2700            command_line    $USER2$/check_postgres.pl -H $HOSTADDRESS$ -u pgsql -db postgres --action locks -w $ARG1$ -c $ARG2$
2701        }
2702
2703
2704        define service {
2705            use                    generic-other
2706            host_name              dbhost.gtld
2707            service_description    dbhost PostgreSQL Service Database Usage Size
2708            check_command          check_postgres_size!256000000!512000000
2709        }
2710
2711        define service {
2712            use                    generic-other
2713            host_name              dbhost.gtld
2714            service_description    dbhost PostgreSQL Service Database Locks
2715            check_command          check_postgres_locks!2!3
2716        }
2717
2719       Copyright (c) 2007-2017 Greg Sabino Mullane <greg@endpoint.com>.
2720
2721       Redistribution and use in source and binary forms, with or without
2722       modification, are permitted provided that the following conditions are
2723       met:
2724
2725         1. Redistributions of source code must retain the above copyright notice,
2726            this list of conditions and the following disclaimer.
2727         2. Redistributions in binary form must reproduce the above copyright notice,
2728            this list of conditions and the following disclaimer in the documentation
2729            and/or other materials provided with the distribution.
2730
2731       THIS SOFTWARE IS PROVIDED BY THE AUTHOR "AS IS" AND ANY EXPRESS OR
2732       IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
2733       WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
2734       DISCLAIMED. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT,
2735       INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
2736       (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
2737       SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
2738       HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
2739       STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING
2740       IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
2741       POSSIBILITY OF SUCH DAMAGE.
2742
2743
2744
2745perl v5.28.1                      2018-05-30                 CHECK_POSTGRES(1)
Impressum