You are viewing a plain text version of this content. The canonical link for it is here.
Posted to reviews@kudu.apache.org by "Alexey Serbin (Code Review)" <ge...@cloudera.org> on 2022/09/01 00:23:59 UTC
[kudu-CR] wip KUDU-2353: tool to parse metrics log
Alexey Serbin has uploaded a new patch set (#10) to the change originally created by Andrew Wong. ( http://gerrit.cloudera.org:8080/12570 )
Change subject: wip KUDU-2353: tool to parse metrics log
......................................................................
wip KUDU-2353: tool to parse metrics log
C++ implementation of the metrics log parser script. There are a couple
functional differences between this tool and the existing script:
- This tool allows filtering metrics by tablet id.
- Histogram metrics for this tool also spit out the rough count of the
measurements.
Note that:
- Kudu metrics are only output when a metric has changed, so this patch tracks
the metric values per entity (per tablet ID or per server) at each
point in time in order to output the correct values. This means that
if within a given set of files, a tablet's metric has not changed, the
tablet will not be accounted for at all by this tool.
- Kudu histogram metrics do spit out a summary for percentiles. We explicitly
do not use this and instead manually generate these from the histogram
counts. While less accurate (IIUC, the counts can be lossy), this
allows us to generate aggregated summaries from multiple tablets.
wip: I added some very basic testing. Could use more.
Here's an example:
[awong@va1022 release]$ ./bin/kudu diagnose parse_metrics kudu-tserver.worker12.foobar.com.kudu.diagnostics.20210123-201217.0.74565 --simple_metrics=tablet.scans_started:num_scans_started --rate_metrics=tablet.scans_started:scans_started_per_sec --histogram_metrics=server.scanner_duration:scanner_duration_
us,server.handler_latency_kudu_tserver_TabletServerService_Scan:scan_rpc_us
I0131 11:53:27.010298 151768 diagnostics_log_parser.cc:272] collecting simple metric tablet.scans_started as num_scanners_started
I0131 11:53:27.010438 151768 diagnostics_log_parser.cc:279] collecting rate metric tablet.scans_started as scanners_started_per_sec
I0131 11:53:27.010455 151768 diagnostics_log_parser.cc:286] collecting histogram metric server.handler_latency_kudu_tserver_TabletServerService_Scan as scan_rpc_us
I0131 11:53:27.010524 151768 diagnostics_log_parser.cc:286] collecting histogram metric server.scanner_duration as scanner_duration_us
timestamp num_scanners_started scanners_started_per_sec scan_rpc_us_count scan_rpc_us_min scan_rpc_us_p50 scan_rpc_us_p75 scan_rpc_us_p95 scan_rpc_us_p99 scan_rpc_us_p99_99 scan_rpc_us_max scanner_duration_us_count scanner_duration_us_min scanner_duration_us_p50 scanner_duration_us_p75 scanner_duration_us_p95 scanner_duration_us_p99 scanner_duration_us_p99_99 scanner_duration_us_max
1611432793767488 68492 0 434170147 2 1215 1639 3711 12927 501759 8650751 1854125 12 23295 1302527 54788095 60030975 60030975 60030975
1611432853767552 231516 2717.0637684653134 434198546 2 1215 1639 3711 12927 501759 8650751 1854200 12 23295 1302527 54788095 60030975 60030975 60030975
1611432913767616 349073 1959.2812434333403 434227285 2 1215 1639 3711 12927 501759 8650751 1854306 12 23295 1302527 54788095 60030975 60030975 60030975
1611432973767689 829597 8008.7235893863 434255021 2 1215 1639 3711 12927 501759 8650751 1854517 12 23295 1302527 54788095 60030975 60030975 60030975
1611433033767772 926516 1615.314432148369 434283184 2 1215 1639 3711 12927 501759 8650751 1854605 12 23295 1302527 54788095 60030975 60030975 60030975
1611433093767841 926626 1.8333312250024245 434309627 2 1215 1639 3711 12927 501759 8650751 1854719 12 23295 1302527 54788095 60030975 60030975 60030975
1611433153767902 960053 557.11610026529809 434339928 2 1215 1639 3711 12927 501759 8650751 1854788 12 23295 1302527 54788095 60030975 60030975 60030975
1611433213767967 1009625 826.19910495096963 434366776 2 1215 1639 3711 12927 501759 8650751 1854831 12 23295 1302527 54788095 60030975 60030975 60030975
1611433273768032 1059960 838.91575784126235 434394555 2 1215 1639 3711 12927 501759 8650751 1854966 12 23295 1302527 54788095 60030975 60030975 60030975
1611433333768067 1061577 26.949984279175837 434420683 2 1215 1639 3711 12927 501759 8650751 1855023 12 23295 1302527 54788095 60030975 60030975 60030975
1611433393768130 1082096 341.98297425121041 434447991 2 1215 1639 3711 12927 501759 8650751 1855185 12 23295 1302527 54788095 60030975 60030975 60030975
1611433453768205 1083102 16.76664570835953 434476348 2 1215 1639 3711 12927 501759 8650751 1855285 12 23295 1302527 54788095 60030975 60030975 60030975
1611433513768270 1088338 87.2665721278802 434498551 2 1215 1639 3711 12927 501759 8650751 1855388 12 23295 1302527 54788095 60030975 60030975 60030975
Change-Id: I8077fb4f6b41fe4b2bd6c877af379ea7a9f415b1
---
M src/kudu/gutil/map-util.h
M src/kudu/tools/diagnostics_log_parser-test.cc
M src/kudu/tools/diagnostics_log_parser.cc
M src/kudu/tools/diagnostics_log_parser.h
M src/kudu/tools/kudu-tool-test.cc
M src/kudu/tools/tool_action_diagnose.cc
A src/kudu/util/regex.h
7 files changed, 988 insertions(+), 310 deletions(-)
git pull ssh://gerrit.cloudera.org:29418/kudu refs/changes/70/12570/10
--
To view, visit http://gerrit.cloudera.org:8080/12570
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings
Gerrit-Project: kudu
Gerrit-Branch: master
Gerrit-MessageType: newpatchset
Gerrit-Change-Id: I8077fb4f6b41fe4b2bd6c877af379ea7a9f415b1
Gerrit-Change-Number: 12570
Gerrit-PatchSet: 10
Gerrit-Owner: Andrew Wong <an...@g.ucla.edu>
Gerrit-Reviewer: Alexey Serbin <al...@apache.org>
Gerrit-Reviewer: Andrew Wong <an...@g.ucla.edu>
Gerrit-Reviewer: Kudu Jenkins (120)
Gerrit-Reviewer: Tidy Bot (241)