Test result ingestion into BigQuery via Kettle is hiccuping increasingly frequently

[spiffxp]

What happened:
http://velodrome.k8s.io/dashboard/db/bigquery-metrics?panelId=12&fullscreen&orgId=1&from=now-6M&to=now

Lots of alerts happening. We've only had one or two weeks in Q2 where an alert hasn't fired.

What you expected to happen:

No alerts to fire.

How to reproduce it (as minimally and precisely as possible):

Please provide links to example occurrences, if any:

Anything else we need to know?:

I suspect this isn't due to Kettle being written in python, but rather the pattern that Kettle follows. It may require a redesign or rewrite but I'm not necessarily sure it's mandated that we switch languages for performance. I suspect i/o is our problem here (due to unbounded growth of a sqlite database)

/area kettle
/area metrics

[spiffxp]

/sig testing

[spiffxp]

/priority important-soon
I am concerned this is going to tip over at some point

[fejta-bot]

Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle stale

[fejta-bot]

Stale issues rot after 30d of inactivity.
Mark the issue as fresh with /remove-lifecycle rotten.
Rotten issues close after an additional 30d of inactivity.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle rotten

[spiffxp]

/remove-lifecycle rotten
It's still hiccuping, but alerts are no longer firing as often

[fejta-bot]

Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle stale

[spiffxp]

/remove-lifecycle stale
/close

[k8s-ci-robot]

@spiffxp: Closing this issue.

In response to this:

/remove-lifecycle stale
/close

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.

[spiffxp]

/reopen

We no longer have velodrome. But we do have a job that fails if k8s-gubernator:build.all is stale: https://testgrid.k8s.io/sig-testing-misc#metrics-kettle&width=5

Most recent example

• at 7:33am PT k8s-gubernator:build.all is >6h old
• at 11:37am PT k8s-gubernator:build.all is still >6h old
• by 12:38pm PT k8s-gubernator:build.all is fresh again

This suggests kettle is pretty consistently not updating bigquery from 1am - 12pm PT, which means go.k8s.io/triage is serving stale data during CET working hours

[k8s-ci-robot]

@spiffxp: Reopened this issue.

In response to this:

/reopen

We no longer have velodrome. But we do have a job that fails if k8s-gubernator:build.all is stale: https://testgrid.k8s.io/sig-testing-misc#metrics-kettle&width=5

Most recent example

• at 7:33am PT k8s-gubernator:build.all is >6h old
• at 11:37am PT k8s-gubernator:build.all is still >6h old
• by 12:38pm PT k8s-gubernator:build.all is fresh again

This suggests kettle is pretty consistently not updating bigquery from 1am - 12pm PT, which means go.k8s.io/triage is serving stale data during CET working hours

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.

[spiffxp]

FYI @MushuEE

[MushuEE]

@spiffxp thanks, will start looking into it.
/assign

[MushuEE]

Span is increasing deltas, not sure if kettle is taking longer to process all the data or not https://k8s-testgrid.appspot.com/sig-testing-misc#metrics-kettle

It might be worth it to start publishing the time that steps take to run within commands. Will add logging shortly

[MushuEE]

 resource: {…}  
 severity: "ERROR"  
 textPayload: "Traceback (most recent call last):
  File "/kettle/update.py", line 59, in <module>
    main()
  File "/kettle/update.py", line 47, in main
    call(bq_cmd + bq_ext + ' k8s-gubernator:build.week build_week.json.gz schema.json')
  File "/kettle/update.py", line 25, in call
    raise OSError('invocation failed')
OSError: invocation failed
" 

Seeing fairly frequent failure on build.week calls. Though there is not much more context...

[fejta-bot]

Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle stale

[MushuEE]

/remove-lifecycle stale

[spiffxp]

Still seeing hiccups, but the duration has slowed

[MushuEE]

Thanks, I will check on the instances

[MushuEE]

Does appear we had another lockup

[fejta-bot]

Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.

If this issue is safe to close now please do so with /close.

Send feedback to sig-contributor-experience at kubernetes/community.
/lifecycle stale

[fejta-bot]

Stale issues rot after 30d of inactivity.
Mark the issue as fresh with /remove-lifecycle rotten.
Rotten issues close after an additional 30d of inactivity.

If this issue is safe to close now please do so with /close.

Send feedback to sig-contributor-experience at kubernetes/community.
/lifecycle rotten

[k8s-triage-robot]

Rotten issues close after 30d of inactivity.
Reopen the issue with /reopen.
Mark the issue as fresh with /remove-lifecycle rotten.

Send feedback to sig-contributor-experience at kubernetes/community.
/close

[k8s-ci-robot]

@k8s-triage-robot: Closing this issue.

In response to this:

Rotten issues close after 30d of inactivity.
Reopen the issue with /reopen.
Mark the issue as fresh with /remove-lifecycle rotten.

Send feedback to sig-contributor-experience at kubernetes/community.
/close

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.