1PMREP(1)                    General Commands Manual                   PMREP(1)
2
3
4

NAME

6       pmrep - performance metrics reporter
7

SYNOPSIS

9       pmrep   [-12357CdgGHIjkLmnprRuUvVxz?]   [-4  action]  [-6  sort-metric]
10       [-8|-9 limit] [-a archive] [-A align]  [--archive-folio  folio]  [-b|-B
11       space-scale]  [-c config] [--container container] [--daemonize] [-e de‐
12       rived] [-E lines] [-f format] [-F outfile]  [-h  host]  [-i  instances]
13       [--include-texts]  [-J  rank]  [-K  spec] [-l delimiter] [-N predicate]
14       [--no-inst-info] [-o  output]  [-O  origin]  [-P|-0  precision]  [-q|-Q
15       count-scale]  [-s  samples]  [-S  starttime] [-t interval] [-T endtime]
16       [-w|-W width] [-X label] [-y|-Y time-scale]  [-Z  timezone]  metricspec
17       [...]
18

DESCRIPTION

20       pmrep is a customizable performance metrics reporting tool.  Any avail‐
21       able performance metric, live or archived, system  and/or  application,
22       can  be  selected  for  reporting  using one of the output alternatives
23       listed below together with applicable formatting options.
24
25       pmrep collects selected metric values through  the  facilities  of  the
26       Performance  Co-Pilot  (PCP),  see  PCPIntro(1).  The metrics to be re‐
27       ported are specified on the command line, in  configuration  files,  or
28       both.   Metrics can be automatically converted and scaled using the PCP
29       facilities, either by default or by per-metric scaling  specifications.
30       In addition to the existing metrics, derived metrics can be defined us‐
31       ing the arithmetic expressions described in pmRegisterDerived(3).
32
33       A wide range of metricsets (see below) is included by default,  provid‐
34       ing  reports  on per-process details, NUMA performance, mimicking other
35       tools like sar(1) and  more,  see  the  pmrep  configuration  files  in
36       $PCP_SYSCONF_DIR/pmrep  (typically  /etc/pcp/pmrep)  for  details.  Tab
37       completion for options, metrics, and metricsets is available  for  bash
38       and zsh.
39
40       Unless  directed  to  another host by the -h option, pmrep will contact
41       the Performance Metrics Collector Daemon (PMCD, see pmcd(1)) on the lo‐
42       cal host.
43
44       The  -a  option  causes  pmrep to use the specified set of archive logs
45       rather than connecting to a PMCD.  The -a and -h options  are  mutually
46       exclusive.
47
48       The  -L  option  causes pmrep to use a local context to collect metrics
49       from DSO PMDAs (Performance Metrics Domain Agents, ``plugins'') on  the
50       local host without PMCD.  Only some metrics are available in this mode.
51       The -a, -h, and -L options are mutually exclusive.
52
53       The metrics of interest are named in the metricspec argument(s).  If  a
54       metricspec  specifies  a  non-leaf node in the Performance Metrics Name
55       Space (PMNS), then pmrep will recursively descend the PMNS  and  report
56       on  all  leaf nodes (i.e., metrics) for that metricspec.  Use pminfo(1)
57       to list all the metrics (PMNS lead nodes) and their descriptions.
58
59       A metricspec has three different forms.  First, on the command line  it
60       can  start with a colon (``:'') to indicate a metricset to be read from
61       pmrep configuration files (see -c and pmrep.conf(5)),  which  may  then
62       consist  of  any number of metrics.  Second, a metricspec starting with
63       non-colon specifies a PMNS node as described above, optionally followed
64       by  metric  output formatting definitions.  This so-called compact form
65       of a metricspec is defined as follows:
66
67     metric[,label[,instances[,unit/scale[,type[,width[,precision[,limit]]]]]]]
68
69       A valid PMNS node (metric) is mandatory.  It may be followed by a  text
70       label  used  with stdout output.  The optional instances definition re‐
71       stricts csv and stdout reporting to the specified instances of the met‐
72       ric  so  non-matching  instances will be filtered out (see -i).  An op‐
73       tional unit/scale is applicable  for  dimension-compatible,  non-string
74       metrics.   See  below  for supported unit/scale specifications.  By de‐
75       fault, cumulative counter metrics are converted to rates,  an  optional
76       type  can  be  set  to raw to disable this rate conversion.  For stdout
77       output a numeric width can be used to set the width of the output  col‐
78       umn  for this metric.  Too wide strings in the output will be truncated
79       to fit the column.  A metric-specific precision can be provided for nu‐
80       meric  non-integer  output values.  Lastly, a metric-specific limit can
81       be set for filtering out numeric values per the limit.
82
83       As a special case  for  metrics  that  are  counters  with  time  units
84       (nanoseconds  to  hours),  the unit/scale can be used to change the de‐
85       fault reporting (for example, milliseconds / second)  to  normalize  to
86       the range zero to one by setting this to sec (see also -y and -Y).
87
88       The  following  metricspec requests the metric kernel.all.sysfork to be
89       reported under the text label forks, converting to the  metric  default
90       rate  count/s  in  an  8 wide column.  Although the definitions in this
91       compact form are optional, they must always be provided  in  the  order
92       specified above, thus the commas.
93
94               kernel.all.sysfork,forks,,,,8
95
96       The  third  form  of a metricspec, verbose form, is described and valid
97       only in pmrep.conf(5).
98
99       Derived metrics are specified like regular PMNS leaf node metrics.
100
101       Options via environment values (see pmGetOptions(3)) override the  cor‐
102       responding  built-in  default  values (if any).  Configuration file op‐
103       tions override the corresponding environment variables (if any).   Com‐
104       mand line options override the corresponding configuration file options
105       (if any).
106

OPTIONS

108       The available command line options are:
109
110       -0 precision, --precision-force=precision
111            Like -P but this option will override per-metric specifications.
112
113       -1, --dynamic-header
114            Print a new dynamically adjusted header every time changes in  the
115            availability  of  metric  and instance values occur.  By default a
116            static header that never changes is printed once.   See  also  -4,
117            -7, and -E.
118
119       -2, --overall-rank
120            Perform overall ranking of instances in archive.  By default rank‐
121            ing (see -J) and reporting happens on each  interval.   With  this
122            option all instances and values are ranked before a summary is re‐
123            ported.  See pmlogsummary(1) for further archive summary reporting
124            alternatives, including averages and peak times for values.
125
126       -3, --overall-rank-alt
127            Like  -2 but print metric instances in pmrep metricspec format, to
128            allow easily selecting the instances for further investigation.
129
130       -4 action, --names-change=action
131            Specify which action to take on receiving a  metric  names  change
132            event  during  sampling.  These events occur when a PMDA discovers
133            new metrics sometime after starting up, and informs running client
134            tools  like  pmrep.   Valid  values for action are update (refresh
135            metrics being sampled), ignore (do nothing -  the  default  behav‐
136            iour)  and  abort (exit the program if such an event occurs).  up‐
137            date implies --dynamic-header.
138
139       -5, --ignore-unknown
140            Silently ignore any metric name that cannot be resolved.  At least
141            one metric must be found for the tool to start.
142
143       -6, --sort-metric=sort-metric
144            Specify  a sort reference metric to sort output by values with -X.
145            By default sorting order is descending, prepending the metric name
146            with the minus sign (``-'') will change the order to be ascending.
147            See also -J and -N.
148
149       -7, --fixed-header
150            With -X print a fixed header once (unless using -E) including  all
151            metrics  being reported.  Unlike with the default (static) header,
152            only instances with values available are  reported.   Unlike  with
153            the  dynamic  header, the header is not updated even if values for
154            some metrics later become (un)available.  See also -1 and -E.
155
156       -8 limit, --limit-filter=limit
157            Limit results to instances with values above/below limit.  A posi‐
158            tive  integer  will  include instances with values at or above the
159            limit in reporting.  A negative  integer  will  include  instances
160            with  values  at or below the limit in reporting.  A value of zero
161            performs no limit filtering.  This option will not override possi‐
162            ble per-metric specifications.  See also -J and -N.
163
164       -9 limit, --limit-filter-force=limit
165            Like -8 but this option will override per-metric specifications.
166
167       -a archive, --archive=archive
168            Performance  metric  values  are retrieved from the set of Perfor‐
169            mance Co-Pilot (PCP) archive log files identified by  the  archive
170            argument,  which is a comma-separated list of names, each of which
171            may be the base name of an archive or the name of a directory con‐
172            taining one or more archives.  See also -u.
173
174       -A align, --align=align
175            Force  the initial sample to be aligned on the boundary of a natu‐
176            ral time unit align.  Refer to PCPIntro(1) for a complete descrip‐
177            tion of the syntax for align.
178
179       --archive-folio=folio
180            Read  metric source archives from the PCP archive folio created by
181            tools like pmchart(1) or, less often, manually with mkaf(1).
182
183       -b scale, --space-scale=scale
184            Unit/scale for  space  (byte)  metrics,  possible  values  include
185            bytes, Kbytes, KB, Mbytes, MB, and so forth.  This option will not
186            override possible per-metric specifications.  See  also  pmParseU‐
187            nitsStr(3).
188
189       -B scale, --space-scale-force=scale
190            Like -b but this option will override per-metric specifications.
191
192       -c config, --config=config
193            Specify  the config file or directory to use.  In case config is a
194            directory all files in it ending .conf will be included.  The  de‐
195            fault  is  the  first  found  of: ./pmrep.conf, $HOME/.pmrep.conf,
196            $HOME/pcp/pmrep.conf,    $PCP_SYSCONF_DIR/pmrep/pmrep.conf,    and
197            $PCP_SYSCONF_DIR/pmrep.  See pmrep.conf(5).
198
199       --container=container
200            Fetch performance metrics from the specified container, either lo‐
201            cal or remote (see -h).
202
203       -C, --check
204            Exit before reporting any values, but after parsing the configura‐
205            tion and metrics and printing possible headers.
206
207       -d, --delay
208            When replaying from an archive, this option requests that the pre‐
209            vailing real-time delay be applied between samples (see -t) to ef‐
210            fect  a  pause,  rather than the default behaviour of replaying at
211            full speed.
212
213       --daemonize
214            Daemonize on startup.
215
216       -e derived, --derived=derived
217            Specify derived performance metrics.  If  derived  starts  with  a
218            slash  (``/'')  or  with a dot (``.'') it will be interpreted as a
219            PCP derived metrics configuration file, otherwise it will  be  in‐
220            terpreted  as comma- or semicolon-separated derived metric expres‐
221            sions.  For complete description of derived metrics  and  PCP  de‐
222            rived  metrics  configuration files see pmLoadDerivedConfig(3) and
223            pmRegisterDerived(3).  Alternatively, using pmrep.conf(5) configu‐
224            ration  syntax  allows defining derived metrics as part of metric‐
225            sets.
226
227       -E lines, --repeat-header=lines
228            Repeat the header every lines of output.  When not using -1 or  -7
229            use  auto to repeat the header based on terminal height.  See also
230            -1 and -7.
231
232       -f format, --timestamp-format=format
233            Use the format string for formatting the  timestamp.   The  format
234            will  be  used  with  Python's  datetime.strftime  method which is
235            mostly the same as that described in strftime(3).  An empty format
236            string (i.e., "") will remove the timestamps from the output.  De‐
237            faults to %H:%M:%S when using the stdout output target.   Defaults
238            to %Y-%m-%d %H:%M:%S when using the csv output target.
239
240       -F outfile, --output-file=outfile
241            Specify the output file outfile.  See -o.
242
243       -g, --separate-header
244            Output the column number and complete metric information, one-per-
245            line, before printing the metric values.
246
247       -G, --no-globals
248            Do not include global metrics in reporting (see pmrep.conf(5)).
249
250       -h host, --host=host
251            Fetch performance metrics from pmcd(1) on host, rather  than  from
252            the default localhost.
253
254       -H, --no-header
255            Do not print any headers.
256
257       -i instances, --instances=instances
258            Retrieve  and  report only the specified metric instances.  By de‐
259            fault all initially present instances are  reported,  except  when
260            writing  an archive (see -o), where also all future instances will
261            be reported.
262
263            The specified instances are filtered from  the  initially  present
264            instances  when  the  tool  is  starting up.  Thus instances which
265            would match the filter appearing after the  tool  has  started  up
266            will not be reported, use -j to change this.
267
268            This  is  a  global option that is used for all set-valued metrics
269            unless a metric-specific instance filter is provided as part of  a
270            metricspec.  By default single-valued ``flat'' metrics without in‐
271            stances are still reported as usual, use -v to change this.
272
273            instances is a comma-separated list of one or more instance filter
274            specifications.   Filters  containing commas or whitespace must be
275            quoted with single (') or double (") quotes.  Note that as part of
276            a metricspec on command line a list with more than one filter both
277            the list and each filter must be quoted as  shown  below.   It  is
278            also  possible to define a single filter with bars (|) as instance
279            separating regex in order to make quoting easier, see below.
280
281            Multiple -i options are allowed as an alternative way of  specify‐
282            ing more than one non-metric-specific instance filters.
283
284            An individual instance filter can be one of the following:
285
286            name      Full  instance  name.  For example, sda for disk.dev in‐
287                      stances or eth0 for network.interface instances.
288
289            PID       Process ID for proc instances.
290
291            command   Base name of a process for proc instances.  For example,
292                      pmcd  would  match  all  pmcd(1) processes regardless of
293                      their path or PID.
294
295            regex     Regular expression.  For example, .*python.* would match
296                      all  instances having the string python as part of their
297                      instance name, meaning that this would match all  Python
298                      processes regardless of their path, PID, or version.
299
300            As an example, the following would report the same instances:
301
302                 $ pmrep -i '. minute' kernel.all.load
303                 $ pmrep -i '1 minute','5 minute' kernel.all.load
304                 $ pmrep -i "'1 minute','5 minute'" kernel.all.load
305                 $ pmrep -i '1 minute' -i '5 minute' kernel.all.load
306                 $ pmrep kernel.all.load,,"'1 minute','5 minute'"
307                 $ pmrep kernel.all.load,,'1 minute|5 minute'
308
309
310            However, this would report only the 1-minute instance:
311
312                 $ pmrep -i '1 minute','5 minute' kernel.all.load,,'1 minute'
313
314
315            And this would report all instances (due to per-metric regex):
316
317                 $ pmrep -i '1 minute','5 minute' 'kernel.all.load,,.*'
318
319
320       -I, --ignore-incompat
321            Ignore  incompatible  metrics.   By  default  incompatible metrics
322            (that is, their type is unsupported or they cannot  be  scaled  as
323            requested)  will  cause  pmrep to terminate with an error message.
324            With this option all incompatible  metrics  are  silently  omitted
325            from  reporting.   This  may  be especially useful when requesting
326            non-leaf nodes of the PMNS tree for reporting.
327
328       --include-texts
329            When writing a PCP archive, include PCP metric help texts  in  the
330            created archive.
331
332       -j, --live-filter
333            Perform  instance live filtering.  This allows capturing all named
334            instances even if processes are restarted at  some  point  (unlike
335            without  live  filtering).   Performing live filtering over a huge
336            number of instances will add some internal overhead so  a  bit  of
337            user caution is advised.  See also -1 and -n.
338
339       -J rank, --rank=rank
340            Limit  results  to  highest/lowest  ranked instances of set-valued
341            metrics.  A positive integer will include highest valued instances
342            in  reporting.   A negative integer will include lowest valued in‐
343            stances in reporting.  A value of zero performs no ranking.  Rank‐
344            ing does not imply sorting, see -6.  See also -2 and -8.
345
346       -k, --extended-csv
347            Write extended CSV output, similar to sadf(1).
348
349       -K spec, --spec-local=spec
350            When fetching metrics from a local context (see -L), the -K option
351            may be used to control the DSO PMDAs that should be made  accessi‐
352            ble.   The  spec  argument conforms to the syntax described in pm‐
353            SpecLocalPMDA(3).  More than one -K option may be used.
354
355       -l delimiter, --delimiter=delimiter
356            Specify the delimiter that separates each column of csv or  stdout
357            output.   The  default for stdout is two spaces (``  '') and comma
358            (``,'') for csv.  When using a non-whitespace delimiter,  all  in‐
359            stances  of the delimiter in string values will be replaced by the
360            underscore (``_'') character.
361
362       -L, --local-PMDA
363            Use a local context to collect metrics from DSO PMDAs on the local
364            host without PMCD.  See also -K.
365
366       -m, --include-labels
367            Include PCP metric labels in the output.
368
369       -n, --invert-filter
370            Perform  ranking  before live filtering.  By default instance live
371            filtering (when requested, see -j) happens before instance ranking
372            (when  requested, see -J).  With this option the logic is inverted
373            and ranking happens before live filtering.
374
375       -N predicate, --predicate=predicate
376            Specify a comma-separated list of predicate filter reference  met‐
377            rics.   By  default ranking (see -J) happens for each metric indi‐
378            vidually.  With predicates, ranking is done only for the specified
379            predicate  metrics.   When  reporting, rest of the metrics sharing
380            the same instance domain (see PCPIntro(1)) as the  predicate  will
381            include  only  the  highest/lowest ranking instances of the corre‐
382            sponding predicate.  Ranking does not imply sorting, see -6.
383
384            So for example, using proc.memory.rss  (resident  memory  size  of
385            process) as the predicate metric together with proc.io.total_bytes
386            and mem.util.used as metrics to be reported,  only  the  processes
387            using  most/least (as per -J) memory will be included when report‐
388            ing total bytes written by processes.  Since  mem.util.used  is  a
389            single-valued metric (thus not sharing the same instance domain as
390            the process related metrics), it will be reported as usual.
391
392       --no-inst-info
393            Omit instance information from headers.  Not applicable with sepa‐
394            rate header (see -g).
395
396       -o output, --output=output
397            Use  output  target  for reporting.  The default target is stdout.
398            The available output target alternatives are:
399
400            archive
401              Record metrics into a PCP archive which can  later  be  replayed
402              with  PCP  tools, including pmrep itself.  See LOGARCHIVE(5) and
403              PCPIntro(1) for details about PCP archive files.  Requires -F.
404
405            csv
406              Print metrics in CSV format (subject to formatting options).
407
408            stdout
409              Print metrics to stdout (format subject to formatting options).
410
411       -O origin, --origin=origin
412            When reporting archived metrics, start reporting at origin  within
413            the  time window (see -S and -T).  Refer to PCPIntro(1) for a com‐
414            plete description of the syntax for origin.
415
416       -p, --timestamps
417            Print timestamps.  By default no timestamps are printed.
418
419       -P precision, --precision=precision
420            Use precision for numeric non-integer output values.  If the value
421            is  too wide for its column width, precision is reduced one by one
422            until the value fits, or not printed at all if it does  not.   The
423            default is to use 3 decimal places (when applicable).  This option
424            will not override possible per-metric specifications.
425
426       -q scale, --count-scale=scale
427            Unit/scale for count metrics,  possible  values  include  count  x
428            10^-1, count, count x 10, count x 10^2, and so forth from 10^-8 to
429            10^7.  (These values are currently space-sensitive.)  This  option
430            will  not  override  possible per-metric specifications.  See also
431            pmParseUnitsStr(3).
432
433       -Q scale, --count-scale-force=scale
434            Like -q but this option will override per-metric specifications.
435
436       -r, --raw
437            Output raw metric values, do not convert  cumulative  counters  to
438            rates.   When  writing archives, raw values are always used.  This
439            option will override possible per-metric specifications.
440
441       -R, --raw-prefer
442            Like -r but this option will not  override  per-metric  specifica‐
443            tions.
444
445       -s samples, --samples=samples
446            The samples argument defines the number of samples to be retrieved
447            and reported.  If samples is 0 or -s is not specified, pmrep  will
448            sample  and  report  continuously (in real time mode) or until the
449            end of the set of PCP archives (in archive mode).  See also -T.
450
451       -S starttime, --start=starttime
452            When reporting archived metrics, the report will be restricted  to
453            those  records logged at or after starttime.  Refer to PCPIntro(1)
454            for a complete description of the syntax for starttime.
455
456       -t interval, --interval=interval
457            Set the reporting interval to something other than the  default  1
458            second.   The  interval  argument  follows the syntax described in
459            PCPIntro(1), and in the simplest form may be an  unsigned  integer
460            (the implied units in this case are seconds).  See also the -T and
461            -u options.
462
463       -T endtime, --finish=endtime
464            When reporting archived metrics, the report will be restricted  to
465            those  records  logged before or at endtime.  Refer to PCPIntro(1)
466            for a complete description of the syntax for endtime.
467
468            When used to define the runtime before pmrep will exit, if no sam‐
469            ples is given (see -s) then the number of reported samples depends
470            on interval (see -t).  If samples is given then interval  will  be
471            adjusted  to  allow  reporting of samples during runtime.  In case
472            all of -T, -s, and -t are given,  endtime  determines  the  actual
473            time pmrep will run.
474
475       -u, --no-interpol
476            When  reporting  archived  metrics, by default values are reported
477            according to the selected sample interval (-t option), not accord‐
478            ing  to  the actual record interval in an archive.  To this effect
479            PCP interpolates the values to be reported based on the records in
480            the  archive.   With the -u option uninterpolated reporting is en‐
481            abled, every recorded value for the selected metrics  is  reported
482            and the requested sample interval (-t) is ignored.
483
484            So  for example, if a PCP archive contains recorded values for ev‐
485            ery 10 seconds and the requested sample interval is 1 hour, by de‐
486            fault pmrep will use an interpolation scheme to compute the values
487            of the requested metrics from the values recorded in the proximity
488            of  these  requested  metrics  and values for every 1 hour are re‐
489            ported.  With -u every record every 10  seconds  are  reported  as
490            such  (the  reported  values are still subject to rate conversion,
491            use -r or -R to disable).
492
493       -U, --no-unit-info
494            Omit unit information from headers.
495
496       -v, --omit-flat
497            Report only set-valued metrics with instances (e.g. disk.dev.read)
498            and  omit  single-valued  ``flat'' metrics without instances (e.g.
499            kernel.all.sysfork).  See -i and -I.
500
501       -V, --version
502            Display version number and exit.
503
504       -w width, --width=width
505            Set the stdout output column width.  Strings will be truncated  to
506            this  width.   The  default width is the shortest that can fit the
507            metric text label, the forced minimum is 3.  This option will  not
508            override possible per-metric specifications.
509
510       -W width, --width-force=width
511            Like -w but this option will override per-metric specifications.
512
513       -x, --extended-header
514            Print extended header.
515
516       -X label, --colxrow=label
517            Swap columns and rows in stdout output, reporting one instance per
518            line, using label as the text label for the instance column.   Use
519            an  empty string ("") to enable swapping without a specific column
520            label.  This change in output allows using grep(1) to  filter  re‐
521            sults or to more closely mimic other tools.  See also -i and -6.
522
523       -y scale, --time-scale=scale
524            Unit/scale  for time metrics, possible values include nanosec, ns,
525            microsec, us, millisec, ms, and so forth up to hour, hr.  This op‐
526            tion  will  not  override possible per-metric specifications.  See
527            also pmParseUnitsStr(3).
528
529       -Y scale, --time-scale-force=scale
530            Like -y but this option will override per-metric specifications.
531
532       -z, --hostzone
533            Use the local timezone of the host that is the source of the  per‐
534            formance  metrics,  as  identified  by either the -h or the -a op‐
535            tions.  The default is to use the timezone of the local host.
536
537       -Z timezone, --timezone=timezone
538            Use timezone for the date and time.  Timezone is in the format  of
539            the environment variable TZ as described in environ(7).  Note that
540            when including a timezone string in output, ISO  8601  -style  UTC
541            offsets are used (so something like -Z EST+5 will become UTC-5).
542
543       -?, --help
544            Display usage message and exit.
545

EXAMPLES

547       The  following  examples use the standard PCP facilities for collecting
548       the metric values, no external utilities are  needed.   The  referenced
549       colon-starting  metricsets are part of the default pmrep configuration.
550       With bash and zsh tab completes available options, metrics, and after a
551       colon metricsets.
552
553       Display network interface metrics on the local host:
554           $ pmrep network.interface.total.bytes
555
556       Display all outgoing network metrics for the wlan0 interface:
557           $ pmrep -i wlan0 -v network.interface.out
558
559       Display the slab total usage (in MB) of two specific slab instances:
560           $ pmrep mem.slabinfo.slabs.total_size,,'kmalloc-4k|xfs_inode',MB
561
562       Display  timestamped vmstat(8) like information using megabytes instead
563       of kilobytes and also include the number of inodes used:
564           $ pmrep -p -B MB :vmstat vfs.inodes.count
565
566       Display per-device disk reads and writes from the  host  server1  using
567       two seconds interval and sadf(1) like CSV output format:
568           $ pmrep -h server1 -t 2s -o csv -k disk.dev.read disk.dev.write
569
570       Display processes using at least 100MB of memory using dynamic headers,
571       additionally use -g to display instance (process) names in full:
572           $ pmrep -b MB --limit-filter 100 --dynamic-header proc.memory.rss
573
574       Display the predefined set of metrics from  the  default  pmrep.conf(5)
575       containing details about I/O requests by current pmlogger process(es):
576           $ pmrep -gp -i pmlogger :proc-io
577
578       Display the three most CPU-using processes:
579           $ pmrep -1gUJ 3 proc.hog.cpu
580
581       Display  sar  -w  and sar -W like information at the same time from the
582       PCP archive ./20150921.09.13 showing values recorded between 3 - 5 PM:
583           $ pmrep -a ./20150921.09.13 -S @15:00 -T @17:00 :sar-w :sar-W
584
585       Record most relevant CPU, memory, and I/O related information about ev‐
586       ery  Java  process on the system, present and future, to an archive ./a
587       on one minute interval at every full minute in a background process:
588           $ pmrep --daemonize -A 1m -t 1m -i '.*java.*' -j -o archive -F ./a \
589               :proc-info :proc-cpu :proc-mem :proc-io
590
591       Record all 389 Directory Server, XFS file system,  and  CPU/memory/disk
592       metrics every five seconds for five minutes to a PCP archive ./a:
593        $ pmrep -t 5s -T 5m -o archive -F ./a ds389 xfs kernel.all.cpu mem disk
594
595       Record process memory and I/O information for those processes which are
596       the three most memory-consuming processes:
597        $ pmrep -o archive -F ./a -J 3 -N proc.memory.rss proc.memory proc.io
598

FILES

600       pmrep.conf
601            pmrep configuration file (see -c)
602
603       $PCP_SYSCONF_DIR/pmrep/*.conf
604            system provided default pmrep configuration files
605

PCP ENVIRONMENT

607       Environment variables with the prefix PCP_ are used to parameterize the
608       file  and  directory names used by PCP.  On each installation, the file
609       /etc/pcp.conf contains the  local  values  for  these  variables.   The
610       $PCP_CONF  variable may be used to specify an alternative configuration
611       file, as described in pcp.conf(5).
612
613       For environment variables affecting PCP tools, see pmGetOptions(3).
614

SEE ALSO

616       PCPIntro(1),  mkaf(1),   pcp(1),   pcp-atop(1),   pcp2elasticsearch(1),
617       pcp2graphite(1),     pcp2influxdb(1),     pcp2json(1),    pcp2spark(1),
618       pcp2xlsx(1), pcp2xml(1), pcp2zabbix(1), pmcd(1), pmchart(1), pmdiff(1),
619       pmdumplog(1),  pmdumptext(1),  pminfo(1), pmiostat(1), pmlogextract(1),
620       pmlogsummary(1),  pmprobe(1),  pmstat(1),  pmval(1),  sadf(1),  sar(1),
621       pmGetOptions(3),  pmLoadDerivedConfig(3),  pmParseUnitsStr(3), pmRegis‐
622       terDerived(3),    pmSpecLocalPMDA(3),    strftime(3),    LOGARCHIVE(5),
623       pcp.conf(5), pmrep.conf(5), PMNS(5), environ(7) and vmstat(8).
624
625
626
627Performance Co-Pilot                  PCP                             PMREP(1)
Impressum