opendevreview | Ian Wienand proposed openstack/openstack-zuul-jobs master: Add build-wheel-cache jobs for CentOS Stream 9 https://review.opendev.org/c/openstack/openstack-zuul-jobs/+/836829 | 00:49 |
---|---|---|
dpawlik | gmann, fungi, dansmith: Hey, I'm back | 06:25 |
*** bhagyashris is now known as bhagyashris|out | 06:26 | |
*** rlandy_ is now known as rlandy | 10:50 | |
*** dviroel|out is now known as dviroel | 11:15 | |
fungi | welcome back, dpawlik! | 11:56 |
fungi | dansmith had some ideas about trending test performance in opensearch, once he's awake/around | 11:57 |
dpawlik | ack | 11:59 |
opendevreview | Merged openstack/openstack-zuul-jobs master: Add build-wheel-cache jobs for CentOS Stream 9 https://review.opendev.org/c/openstack/openstack-zuul-jobs/+/836829 | 12:16 |
opendevreview | Merged openstack/project-config master: Build and publish wheel mirror for CentOS Stream 9 https://review.opendev.org/c/openstack/project-config/+/836793 | 12:34 |
*** ralonsoh_ is now known as ralonsoh | 12:34 | |
opendevreview | Merged openstack/ci-log-processing master: Execute vacuum command after cleaning up DB https://review.opendev.org/c/openstack/ci-log-processing/+/834828 | 13:06 |
*** dasm|off is now known as dasm | 13:27 | |
dansmith | dpawlik: yeah, this: https://zuul.opendev.org/t/openstack/build/603ceac5a8c246d3afecc92a5c00d319/log/job-output.txt#26555 | 13:31 |
dansmith | I need to get it generating to a file that we collect in logs obviously | 13:32 |
dpawlik | dansmith: eeem, you want to check logs in opensearch for that build? | 13:34 |
dpawlik | https://opensearch.logs.openstack.org/_dashboards/app/discover?security_tenant=global#/?_g=(filters:!(),refreshInterval:(pause:!t,value:0),time:(from:now-4d,to:now))&_a=(columns:!(_source),filters:!(),index:'94869730-aea8-11ec-9e6a-83741af3fdcd',interval:auto,query:(language:kuery,query:'build_uuid:%22603ceac5a8c246d3afecc92a5c00d319%22'),sort:!()) | 13:34 |
dansmith | dpawlik: no, I'm hoping to get lots of jobs to emit that json blob in a way that we can plot those statistics over (short intervals of) time | 13:37 |
* dpawlik reading conversation from 8 April | 13:38 | |
dpawlik | dansmith: aha, so you would like to get information about utilization for the job to make some statistics, right? | 13:41 |
dansmith | dpawlik: well, once lots of jobs are emitting this, I'd like to be able to compare a given run to to average (for that named job), | 13:42 |
dansmith | and I think we'd also like to feed that into graphite or something similar for longer-term trends, if possible | 13:42 |
dpawlik | dansmith: hmm, it can be complex visualization in Opensearch, but I can try to do it | 13:45 |
dansmith | dpawlik: oh clarkb thought there would be some support for that | 13:45 |
dansmith | however, even if I can just use it to get a bunch of data (like all the memory values for a given job for the past week) that would be useful to calculate the average | 13:46 |
dpawlik | dansmith, understand. Will try to create such visualization | 13:47 |
dansmith | thanks! | 13:47 |
dpawlik | good that you remind me that I need to implement export/import objects from opensearch dashboards and keep it in ci-log-processing repo | 13:49 |
opendevreview | Merged openstack/pbr master: Future-proofing pyproject.toml https://review.opendev.org/c/openstack/pbr/+/835249 | 14:00 |
dpawlik | dansmith: it can be very complex. Need to check if such information can be provided by zuul or it require to create new label... Maybe such information can be provided should be moved to some other file | 14:12 |
dpawlik | dansmith: I will think how to provide such info | 14:12 |
dansmith | dpawlik: I'm working to get this into a .json file in the logs directory instead of just inline in the console.. is that what you mean? | 14:13 |
dpawlik | doing queries with opensearch client via python script seems to be easiest way | 14:13 |
dpawlik | dansmith: yes, this would be nice | 14:14 |
dansmith | dpawlik: yes for sure, I just dumped to the console there for debugging, but it's more useful to everyone to have it in a file of course | 14:14 |
fzzf[m] | clarkb: fungi hi, when build image have follow error , I try add block-device-efi element, it didn't work.... (full message at https://matrix.org/_matrix/media/r0/download/matrix.org/TdeLHPZDzfvMFhOIaYXOPRBF) | 15:15 |
*** dviroel is now known as dviroel|lunch|afk | 15:32 | |
clarkb | dpawlik: dansmith we can do the visiaulize in grafana not kibana | 16:04 |
clarkb | since grafana can use elasticsearch as a data source | 16:04 |
dansmith | the visualization is really less important to me than the calculation of the running average | 16:05 |
clarkb | good point | 16:05 |
dansmith | a visualization over seven days is not very likely to be useful for this kind of thing, it's more like "this one patch is 30% higher than the average for $stat" | 16:06 |
clarkb | ++ | 16:06 |
clarkb | unrelated I tried to drive over 26 thi smorning to drop my parents off at the airport. They missed their flight but the car didn't take a tree to the face or another car to the face and we are warm so win! | 16:06 |
dansmith | heh, yeah this is kinda crazy | 16:06 |
dansmith | we have over seven inches here and it's still coming down | 16:07 |
clarkb | wow. We got to the top of sylvan and drove parralel to the the free way only to get to the onramp and be told "you have to turn around because trees are falling on cars and the freeway is completely closed | 16:07 |
clarkb | this was after playing portland snow driving in the west hills | 16:07 |
clarkb | we took that as a good sign to turn around go home and rebook the flight | 16:07 |
dansmith | yeah | 16:07 |
dansmith | we have nowhere we need to be until later this week | 16:08 |
*** rlandy_ is now known as rlandy | 19:28 | |
*** dmellado_ is now known as dmellado | 21:59 | |
*** dviroel is now known as dviroel|out | 22:02 | |
*** dasm is now known as dasm|off | 23:57 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!