1PCP2ELASTICSEARCH(1) General Commands Manual PCP2ELASTICSEARCH(1)
2
6 pcp2elasticsearch - pcp-to-elasticsearch metrics exporter
7
9 pcp2elasticsearch [-5CGHIjLmnrRvV?] [-4 action] [-8|-9 limit] [-a ar‐
10 chive] [-A align] [--archive-folio folio] [-b|-B space-scale] [-c con‐
11 fig] [--container container] [--daemonize] [-e derived] [-g URL] [-h
12 host] [-i instances] [-J rank] [-K spec] [-N predicate] [-O origin] [-p
13 type] [-P|-0 precision] [-q|-Q count-scale] [-s samples] [-S starttime]
14 [-t interval] [-T endtime] [-x index] [-X hostid] [-y|-Y time-scale]
15 metricspec [...]
16
18 pcp2elasticsearch is a customizable performance metrics exporter tool
19 from PCP to Elasticsearch. Any available performance metric, live or
20 archived, system and/or application, can be selected for exporting us‐
21 ing either command line arguments or a configuration file.
22 pcp2elasticsearch is a close relative of pmrep(1). Refer to pmrep(1)
24 for the metricspec description accepted on pcp2elasticsearch command
25 line. See pmrep.conf(5) for description of the pcp2elasticsearch.conf
26 configuration file overall syntax. This page describes pcp2elastic‐
27 search specific options and configuration file differences with pm‐
28 rep.conf(5). pmrep(1) also lists some usage examples of which most are
29 applicable with pcp2elasticsearch as well.
30 Only the command line options listed on this page are supported, other
32 options recognized by pmrep(1) are not supported.
33 Options via environment values (see pmGetOptions(3)) override the cor‐
35 responding built-in default values (if any). Configuration file op‐
36 tions override the corresponding environment variables (if any). Com‐
37 mand line options override the corresponding configuration file options
38 (if any).
39
41 pcp2elasticsearch uses a configuration file with overall syntax de‐
42 scribed in pmrep.conf(5). The following options are common with pm‐
43 rep.conf: version, source, speclocal, derived, header, globals, sam‐
44 ples, interval, type, type_prefer, ignore_incompat, names_change, in‐
45 stances, live_filter, rank, limit_filter, limit_filter_force, in‐
46 vert_filter, predicate, omit_flat, include_labels, precision, preci‐
47 sion_force, count_scale, count_scale_force, space_scale,
48 space_scale_force, time_scale, time_scale_force. The output option is
49 recognized but ignored for pmrep.conf compatibility.
50 pcp2elasticsearch specific options
52 es_server (string)
53 Specify the URL of Elasticsearch server to send the metrics to.
54 Corresponding command line option is -g. Default is http://local‐
55 host:9200/.
56 es_index (string)
58 Specify the Elasticsearch index for metric names. Corresponding
59 command line option is -x. Default is pcp.
60 es_hostid (string)
62 Specify the Elasticsearch host-id for measurements. Corresponding
63 command line option is -X. Defaults to the metrics source host.
64 es_search_type (string)
66 Specify the Elasticsearch search type for measurements. Corre‐
67 sponding command line option is -p. Defaults to pcp-metric.
68
70 The available command line options are:
71 -0 precision, --precision-force=precision
73 Like -P but this option will override per-metric specifications.
74 -4 action, --names-change=action
76 Specify which action to take on receiving a metric names change
77 event during sampling. These events occur when a PMDA discovers
78 new metrics sometime after starting up, and informs running client
79 tools like pcp2elasticsearch. Valid values for action are update
80 (refresh metrics being sampled), ignore (do nothing - the default
81 behaviour) and abort (exit the program if such an event happens).
82 -5, --ignore-unknown
84 Silently ignore any metric name that cannot be resolved. At least
85 one metric must be found for the tool to start.
86 -8 limit, --limit-filter=limit
88 Limit results to instances with values above/below limit. A posi‐
89 tive integer will include instances with values at or above the
90 limit in reporting. A negative integer will include instances
91 with values at or below the limit in reporting. A value of zero
92 performs no limit filtering. This option will not override possi‐
93 ble per-metric specifications. See also -J and -N.
94 -9 limit, --limit-filter-force=limit
96 Like -8 but this option will override per-metric specifications.
97 -a archive, --archive=archive
99 Performance metric values are retrieved from the set of Perfor‐
100 mance Co-Pilot (PCP) archive log files identified by the archive
101 argument, which is a comma-separated list of names, each of which
102 may be the base name of an archive or the name of a directory con‐
103 taining one or more archives.
104 -A align, --align=align
106 Force the initial sample to be aligned on the boundary of a natu‐
107 ral time unit align. Refer to PCPIntro(1) for a complete descrip‐
108 tion of the syntax for align.
109 --archive-folio=folio
111 Read metric source archives from the PCP archive folio created by
112 tools like pmchart(1) or, less often, manually with mkaf(1).
113 -b scale, --space-scale=scale
115 Unit/scale for space (byte) metrics, possible values include
116 bytes, Kbytes, KB, Mbytes, MB, and so forth. This option will not
117 override possible per-metric specifications. See also pmParseU‐
118 nitsStr(3).
119 -B scale, --space-scale-force=scale
121 Like -b but this option will override per-metric specifications.
122 -c config, --config=config
124 Specify the config file or directory to use. In case config is a
125 directory all files under it ending .conf will be included. The
126 default is the first found of: ./pcp2elasticsearch.conf,
127 $HOME/.pcp2elasticsearch.conf, $HOME/pcp/pcp2elasticsearch.conf,
128 and $PCP_SYSCONF_DIR/pcp2elasticsearch.conf. For details, see the
129 above section and pmrep.conf(5).
130 --container=container
132 Fetch performance metrics from the specified container, either lo‐
133 cal or remote (see -h).
134 -C, --check
136 Exit before reporting any values, but after parsing the configura‐
137 tion and metrics and printing possible headers.
138 --daemonize
140 Daemonize on startup.
141 -e derived, --derived=derived
143 Specify derived performance metrics. If derived starts with a
144 slash (``/'') or with a dot (``.'') it will be interpreted as a
145 derived metrics configuration file, otherwise it will be inter‐
146 preted as comma- or semicolon-separated derived metric expres‐
147 sions. For details see pmLoadDerivedConfig(3) and pmRegister‐
148 Derived(3).
149 -g URL, --es-host=URL
151 Elasticsearch server URL to send the metrics to.
152 -G, --no-globals
154 Do not include global metrics in reporting (see pmrep.conf(5)).
155 -h host, --host=host
157 Fetch performance metrics from pmcd(1) on host, rather than from
158 the default localhost.
159 -H, --no-header
161 Do not print any headers.
162 -i instances, --instances=instances
164 Retrieve and report only the specified metric instances. By de‐
165 fault all instances, present and future, are reported.
166 Refer to pmrep(1) for complete description of this option.
168 -I, --ignore-incompat
170 Ignore incompatible metrics. By default incompatible metrics
171 (that is, their type is unsupported or they cannot be scaled as
172 requested) will cause pcp2elasticsearch to terminate with an error
173 message. With this option all incompatible metrics are silently
174 omitted from reporting. This may be especially useful when re‐
175 questing non-leaf nodes of the PMNS tree for reporting.
176 -j, --live-filter
178 Perform instance live filtering. This allows capturing all named
179 instances even if processes are restarted at some point (unlike
180 without live filtering). Performing live filtering over a huge
181 number of instances will add some internal overhead so a bit of
182 user caution is advised. See also -n.
183 -J rank, --rank=rank
185 Limit results to highest/lowest ranked instances of set-valued
186 metrics. A positive integer will include highest valued instances
187 in reporting. A negative integer will include lowest valued in‐
188 stances in reporting. A value of zero performs no ranking. Rank‐
189 ing does not imply sorting, see -6. See also -8.
190 -K spec, --spec-local=spec
192 When fetching metrics from a local context (see -L), the -K option
193 may be used to control the DSO PMDAs that should be made accessi‐
194 ble. The spec argument conforms to the syntax described in pm‐
195 SpecLocalPMDA(3). More than one -K option may be used.
196 -L, --local-PMDA
198 Use a local context to collect metrics from DSO PMDAs on the local
199 host without PMCD. See also -K.
200 -m, --include-labels
202 Include metric labels in the output.
203 -n, --invert-filter
205 Perform ranking before live filtering. By default instance live
206 filtering (when requested, see -j) happens before instance ranking
207 (when requested, see -J). With this option the logic is inverted
208 and ranking happens before live filtering.
209 -N predicate, --predicate=predicate
211 Specify a comma-separated list of predicate filter reference met‐
212 rics. By default ranking (see -J) happens for each metric indi‐
213 vidually. With predicates, ranking is done only for the specified
214 predicate metrics. When reporting, rest of the metrics sharing
215 the same instance domain (see PCPIntro(1)) as the predicate will
216 include only the highest/lowest ranking instances of the corre‐
217 sponding predicate. Ranking does not imply sorting, see -6.
218 So for example, using proc.memory.rss (resident memory size of
220 process) as the predicate metric together with proc.io.total_bytes
221 and mem.util.used as metrics to be reported, only the processes
222 using most/least (as per -J) memory will be included when report‐
223 ing total bytes written by processes. Since mem.util.used is a
224 single-valued metric (thus not sharing the same instance domain as
225 the process related metrics), it will be reported as usual.
226 -O origin, --origin=origin
228 When reporting archived metrics, start reporting at origin within
229 the time window (see -S and -T). Refer to PCPIntro(1) for a com‐
230 plete description of the syntax for origin.
231 -p type, --es-search-type=type
233 Elasticsearch search type for measurements.
234 -P precision, --precision=precision
236 Use precision for numeric non-integer output values. The default
237 is to use 3 decimal places (when applicable). This option will
238 not override possible per-metric specifications.
239 -q scale, --count-scale=scale
241 Unit/scale for count metrics, possible values include count x
242 10^-1, count, count x 10, count x 10^2, and so forth from 10^-8 to
243 10^7. (These values are currently space-sensitive.) This option
244 will not override possible per-metric specifications. See also
245 pmParseUnitsStr(3).
246 -Q scale, --count-scale-force=scale
248 Like -q but this option will override per-metric specifications.
249 -r, --raw
251 Output raw metric values, do not convert cumulative counters to
252 rates. This option will override possible per-metric specifica‐
253 tions.
254 -R, --raw-prefer
256 Like -r but this option will not override per-metric specifica‐
257 tions.
258 -s samples, --samples=samples
260 The samples argument defines the number of samples to be retrieved
261 and reported. If samples is 0 or -s is not specified, pcp2elas‐
262 ticsearch will sample and report continuously (in real time mode)
263 or until the end of the set of PCP archives (in archive mode).
264 See also -T.
265 -S starttime, --start=starttime
267 When reporting archived metrics, the report will be restricted to
268 those records logged at or after starttime. Refer to PCPIntro(1)
269 for a complete description of the syntax for starttime.
270 -t interval, --interval=interval
272 Set the reporting interval to something other than the default 1
273 second. The interval argument follows the syntax described in
274 PCPIntro(1), and in the simplest form may be an unsigned integer
275 (the implied units in this case are seconds). See also the -T op‐
276 tion.
277 -T endtime, --finish=endtime
279 When reporting archived metrics, the report will be restricted to
280 those records logged before or at endtime. Refer to PCPIntro(1)
281 for a complete description of the syntax for endtime.
282 When used to define the runtime before pcp2elasticsearch will
284 exit, if no samples is given (see -s) then the number of reported
285 samples depends on interval (see -t). If samples is given then
286 interval will be adjusted to allow reporting of samples during
287 runtime. In case all of -T, -s, and -t are given, endtime deter‐
288 mines the actual time pcp2elasticsearch will run.
289 -v, --omit-flat
291 Report only set-valued metrics with instances (e.g. disk.dev.read)
292 and omit single-valued ``flat'' metrics without instances (e.g.
293 kernel.all.sysfork). See -i and -I.
294 -V, --version
296 Display version number and exit.
297 -x index, --es-index=index
299 Elasticsearch index for metric names.
300 -X hostid, --es-hostid=hostid
302 Elasticsearch hostid for measurements.
303 -y scale, --time-scale=scale
305 Unit/scale for time metrics, possible values include nanosec, ns,
306 microsec, us, millisec, ms, and so forth up to hour, hr. This op‐
307 tion will not override possible per-metric specifications. See
308 also pmParseUnitsStr(3).
309 -Y scale, --time-scale-force=scale
311 Like -y but this option will override per-metric specifications.
312 -?, --help
314 Display usage message and exit.
315
317 First, pcp2elasticsearch sends this type definition to Elasticsearch:
318 {
320 "mappings": {
321 "pcp-metric": {
322 "properties": {
323 "@timestamp": {
324 "type": "date"
325 },
326 "host-id": {
327 "type": "string"
328 }
329 }
330 }
331 }
332 }
333 The actual metrics data document is similar to this example:
335 {
337 "@host-id": "bozo",
338 "mem": {
339 "util": {
340 "used": 4525820,
341 "free": 3344000
342 }
343 },
344 "@timestamp": 1507637520,
345 "disk": {
346 "partitions": {
347 "@instances": [
348 {
349 "read": 1121,
350 "@id": "sda1"
351 },
352 {
353 "read": 77306,
354 "@id": "sda2"
355 },
356 {
357 "read": 131,
358 "@id": "sda3"
359 }
360 ]
361 }
362 }
363 }
364
366 pcp2elasticsearch.conf
367 pcp2elasticsearch configuration file (see -c)
368
370 Environment variables with the prefix PCP_ are used to parameterize the
371 file and directory names used by PCP. On each installation, the file
372 /etc/pcp.conf contains the local values for these variables. The
373 $PCP_CONF variable may be used to specify an alternative configuration
374 file, as described in pcp.conf(5).
375 For environment variables affecting PCP tools, see pmGetOptions(3).
377
379 mkaf(1), PCPIntro(1), pcp(1), pcp2graphite(1), pcp2influxdb(1),
380 pcp2json(1), pcp2spark(1), pcp2xlsx(1), pcp2xml(1), pcp2zabbix(1),
381 pmcd(1), pminfo(1), pmrep(1), pmGetOptions(3), pmSpecLocalPMDA(3), pm‐
382 LoadDerivedConfig(3), pmParseUnitsStr(3), pmRegisterDerived(3), LOGA‐
383 RCHIVE(5), pcp.conf(5), PMNS(5) and pmrep.conf(5).
384Performance Co-Pilot PCP PCP2ELASTICSEARCH(1)