NETOBSERV-1247: using Meta informer for replicaSets

[msherif1234]

This is follow up to the work @jpinsonneau started in https://github.com/jpinsonneau/flowlogs-pipeline/tree/update_metainformer

• switch to metadata informer for RS

[openshift-ci-robot]

@msherif1234: This pull request references NETOBSERV-1247 which is a valid jira issue.

In response to this:

This is follow up to the work @jpinsonneau started in https://github.com/jpinsonneau/flowlogs-pipeline/tree/update_metainformer

• switch to metadata informer for RS
• update go-client to v0.28.0

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.

[openshift-ci-robot]

@msherif1234: This pull request references NETOBSERV-1247 which is a valid jira issue.

In response to this:

This is follow up to the work @jpinsonneau started in https://github.com/jpinsonneau/flowlogs-pipeline/tree/update_metainformer

• switch to metadata informer for RS
• update go-client to v0.28.0 and its dependencies

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.

[msherif1234]

/ok-to-test

[github-actions]

New image:
quay.io/netobserv/flowlogs-pipeline:27c0dcf

It will expire after two weeks.

To deploy this build, run from the operator repo, assuming the operator is running:

USER=netobserv VERSION=27c0dcf make set-flp-image

[codecov]

Codecov Report

Attention: 26 lines in your changes are missing coverage. Please review.

Comparison is base (e3a97f1) 67.07% compared to head (aaf2505) 66.90%.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #474      +/-   ##
==========================================
- Coverage   67.07%   66.90%   -0.17%     
==========================================
  Files          94       94              
  Lines        6724     6741      +17     
==========================================
  Hits         4510     4510              
- Misses       1948     1965      +17     
  Partials      266      266              

Flag	Coverage Δ
unittests	66.90% <0.00%> (-0.17%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

Files	Coverage Δ
pkg/pipeline/transform/kubernetes/kubernetes.go	17.94% <0.00%> (-1.20%)	⬇️

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[msherif1234]

/ok-to-test

[openshift-ci-robot]

@msherif1234: This pull request references NETOBSERV-1247 which is a valid jira issue.

In response to this:

This is follow up to the work @jpinsonneau started in https://github.com/jpinsonneau/flowlogs-pipeline/tree/update_metainformer

• switch to metadata informer for RS

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.

[github-actions]

New image:
quay.io/netobserv/flowlogs-pipeline:c6e2afd

It will expire after two weeks.

To deploy this build, run from the operator repo, assuming the operator is running:

USER=netobserv VERSION=c6e2afd make set-flp-image

[msherif1234]

[jpinsonneau]

The initial issue was on a cluster having 15k+ replicasets. We should try to reproduce something around this number 😸

[jpinsonneau]

Code looks good, not tested yet
Do you see differences in the enriched flows ?

[msherif1234]

• using hey-ho tool /hey-ho.sh -n 50, which generate the following RS load

$ oc get rs -A | wc -l
343

if we ignore few flp pods with PR showing much higher usage then I would say this is mostly the same , I am not able to increase RS to lagre number w/o increasing current default of 250 pods/node which prevented me from swapping FLP images to test before and after

[jpinsonneau]

if we ignore few flp pods with PR showing much higher usage then I would say this is mostly the same , I am not able to increase RS to lagre number w/o increasing current default of 250 pods/node which prevented me from swapping FLP images to test before and after

So do you think it's not worth it or should it be verified by QEs ?

[msherif1234]

if we ignore few flp pods with PR showing much higher usage then I would say this is mostly the same , I am not able to increase RS to lagre number w/o increasing current default of 250 pods/node which prevented me from swapping FLP images to test before and after

So do you think it's not worth it or should it be verified by QEs ?

I think QE verified on scale run we might have better numbers there @memodi WDYT ?

[memodi]

I guess the idea here to rollout the new revisions to existing deployments, we don't have any tool that does that today. Simply, increasing the cluster size won't help since FLP pods would scale equally.

[memodi]

@msherif1234 @jpinsonneau - does all FLP tries to build similar cache? I am trying to think if we had say 120 node cluster with 100 deployments on each worker would be appropriate scenario? cc @nathan-weinberg

[jpinsonneau]

@msherif1234 @jpinsonneau - does all FLP tries to build similar cache?

Yes every FLP build the exact same cache. That's why the kube API is taking a burst on Netobserv startup
https://issues.redhat.com/browse/NETOBSERV-1248

I am trying to think if we had say 120 node cluster with 100 deployments on each worker would be appropriate scenario? cc @nathan-weinberg

So you would be around 12k pods ? More if you plan multiple replicas ?
That sounds good ! Compare FLPs overall CPU / Memory without that PR and with that PR.
A plus would also be checking kube api perfs.

[memodi]

/rm-ok-test

[nathan-weinberg]

/ok-to-test

[github-actions]

New image:
quay.io/netobserv/flowlogs-pipeline:89ba22d

It will expire after two weeks.

To deploy this build, run from the operator repo, assuming the operator is running:

USER=netobserv VERSION=89ba22d make set-flp-image

[jotak]

@msherif1234 hope you don"t mind :) I pushed a fix on this PR, FLP was CLBO due to a wrong conversion, the transform func now expecting partial metadata rather than replicaset

[jotak]

/lgtm
For test failing: this should fix it: #534

[github-actions]

New image:
quay.io/netobserv/flowlogs-pipeline:217cdb2

It will expire after two weeks.

To deploy this build, run from the operator repo, assuming the operator is running:

USER=netobserv VERSION=217cdb2 make set-flp-image

[msherif1234]

/ok-to-test

[github-actions]

New image:
quay.io/netobserv/flowlogs-pipeline:e98f7ba

It will expire after two weeks.

To deploy this build, run from the operator repo, assuming the operator is running:

USER=netobserv VERSION=e98f7ba make set-flp-image

[msherif1234]

/hold cancel

[nathan-weinberg]

/label qe-approved

[openshift-ci-robot]

@msherif1234: This pull request references NETOBSERV-1247 which is a valid jira issue.

Warning: The referenced jira issue has an invalid target version for the target branch this PR targets: expected the story to target the "4.15.0" version, but no target version was set.

In response to this:

This is follow up to the work @jpinsonneau started in https://github.com/jpinsonneau/flowlogs-pipeline/tree/update_metainformer

• switch to metadata informer for RS

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.

[msherif1234]

/approve

[openshift-ci]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: msherif1234

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

• OWNERS [msherif1234]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment