Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

PLNSRVCE-1692: consolidate metrics into single pipelinerun, taskrun, pod controllers tuned for performance; convert pvc metric to poll only #62

Merged

Conversation

gabemontero
Copy link
Collaborator

As part of getting ready to add the 'no pod create attempts occuring for pipelineruns' metric for satisfying AppSRE/Infra feedback on core tekton health, I've been monitoring the memory usage on prod-r01 and the previous "simpler" approach of creating controllers per replicas may be now proving untenable. As konflux onboading increases, exporter's memory usage continues to grow, and I've had to increase the memory limit from 1G to 3G just in the last few weeks.

Note, the consolidating metrics into common reconcilers made the PVC gauge metric utilizing a reconciler in addition to the existing poller (a la what upstream does) untenable. Hence we now just leverage the polling thread for the PVC metric.

rh-pre-commit.version: 2.2.0
rh-pre-commit.check-secrets: ENABLED

@enarha @jkhelil PTAL

@jkhelil your github ID does not show up as a reviewer. My guess is there is some github related onboarding activity you still have to do, though I'm not sure what it is exactly. See if you can find it in the onboarding doc, or maybe see if Romain knows.

… tuned for performance; convert pvc metric to poll only; more clean up and reorg

As part of getting ready to add the 'no pod create attempts occuring for pipelineruns' metric for satisfying AppSRE/Infra feedback on core tekton health,
I've been monitoring the memory usage on prod-r01 and the previous "simpler" approach of creating controllers per replicas may be now proving untenable.
As konflux onboading increases, exporter's memory usage continues to grow, and I've had to increase the memory limit from 1G to 3G just in the last few weeks.

Note, the consolidating metrics into common reconcilers made the PVC gauge metric utilizing a reconciler in addition to the existing poller (a la what upstream does)
untenable.  Hence we now just leverage the polling thread for the PVC metric.

rh-pre-commit.version: 2.2.0
rh-pre-commit.check-secrets: ENABLED
rh-pre-commit.version: 2.2.0
rh-pre-commit.check-secrets: ENABLED
@gabemontero gabemontero requested a review from enarha March 29, 2024 17:27
@gabemontero gabemontero changed the title PLNSRVCE-1692: consolidate metrics into single pipelinerun, taskrun, pod controllers tuned for performance; convert pvc metric to poll only; more clean up and reorg PLNSRVCE-1692: consolidate metrics into single pipelinerun, taskrun, pod controllers tuned for performance; convert pvc metric to poll only Mar 29, 2024
@gabemontero
Copy link
Collaborator Author

@jkhelil - Romain gave me admin perms and I've added you to the repo. You're now listed as an official reviewer on this PR

Copy link
Collaborator

@enarha enarha left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

/lgtm

@gabemontero gabemontero merged commit 2a5d242 into openshift-pipelines:main Apr 2, 2024
2 checks passed
@gabemontero gabemontero deleted the consolidate-watches branch April 2, 2024 19:54
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

None yet

3 participants