Decrease memory usage by K8S Clients #2463
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
jdn5126
changed the title
|
Integration test run: https://github.com/aws/amazon-vpc-cni-k8s/actions/runs/5522624871 |
jdn5126
added a commit
that referenced
this pull request
Jul 11, 2023
* refactor canary test to access images from AWS registries (#2398) * upgrade client-go and controller-runtime modules (#2396) * updates for v1.13.0 release (#2400) * chore: Added dependabot (#2403) * dependency updates (#2412) * deprecate ENABLE_NFTABLES and set iptables mode using iptables-wrapper script (#2402) * update networking test agent to go1.20 and latest sys module (#2413) * skip delete test cluster to debug (#2414) * Revert "skip delete test cluster to debug (#2414)" (#2415) This reverts commit 7c30943. * authenticate to test image registry (#2417) * update test agent image (#2419) * update test agent hash in go.mod (#2422) * fix hard-coded nitro instances (#2428) * move authentication step from test canary script (#2429) * node initialization must come after primary ENI's security groups are synced to cache (#2427) * Add 1.27 to Rec Version Table (#2404) * revise rec version table * make DOCKER_ARGS a passable var from CLI builds (#2434) Signed-off-by: jonahjon <jonahjones094@gmail.com> * Update Kops cluster to latest and add parameter for kops version (#2435) * Updates instance limits including c7gn (#2438) * Update Kops cluster to latest and add parameter for kops version (#2440) * update image tag to v1.13.2 (#2432) * update docs and CNI logging (#2433) * remove default canary test run from integration tests (#2443) * Silences nightly cron jobs for forks (#2444) * Silences weekly cron jobs for forks (#2459) * refactor performance tests (#2455) * add custom-networking test covering ENIConfig objects with no security (#2445) groups * k8s clients only need to access corev1; add pod selector (#2463) --------- Signed-off-by: jonahjon <jonahjones094@gmail.com> Co-authored-by: Olivia Song <sonyingy@amazon.com> Co-authored-by: Ellis Tarn <ellistarn@gmail.com> Co-authored-by: Geoffrey Cline <geoffreyc@outlook.com> Co-authored-by: Jonah Jones <jonahjones094@gmail.com> Co-authored-by: Jay Deokar <23660509+jaydeokar@users.noreply.github.com> Co-authored-by: Matt <matt.merkes@gmail.com> Co-authored-by: Matt <merkes@amazon.com>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
What type of PR is this?
bug
Which issue does this PR fix:
#2436
What does this PR do / Why do we need it:
This PR modifies the schemas used by the raw and cached Kubernetes clients in the IPAMD and Metrics Helper agents. These clients only need to add the
corev1scheme: https://github.com/kubernetes/api/blob/v0.26.5/core/v1/register.go#L45 , as the clients only make GET and LIST calls for objects added by this scheme.Decreasing the number of objects loaded decreases the memory consumed by these clients. For IPAMD, a selector is added for pods such that only pods on the deployed node are cached. Metrics Helper does not add this selector, as it needs to query
aws-nodepods on all nodes. Metrics Helper already has a label selector in its pod watcher.If an issue # is not available please add repro steps and logs from IPAMD/CNI showing the issue:
Testing done on this change:
Running the performance tests yielded the following memory decrease:
v1.13.2:this PR:I also validated that all integration tests pass. I will schedule a manual run against this PR as well.
Automation added to e2e:
N/A
Will this PR introduce any new dependencies?:
No
Will this break upgrades or downgrades. Has updating a running cluster been tested?:
No, Yes
Does this change require updates to the CNI daemonset config files to work?:
No
Does this PR introduce any user-facing change?:
Yes
By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.