KEP-6053: Kubelet Image Garbage Collection Observability#6054
KEP-6053: Kubelet Image Garbage Collection Observability#6054mattsu2020 wants to merge 1 commit intokubernetes:masterfrom
Conversation
|
[APPROVALNOTIFIER] This PR is NOT APPROVED This pull-request has been approved by: mattsu2020 The full list of commands accepted by this bot can be found here. DetailsNeeds approval from an approver in each of these files:Approvers can indicate their approval by writing |
|
Hi @mattsu2020. Thanks for your PR. I'm waiting for a kubernetes member to verify that this patch is reasonable to test. If it is, they should reply with Regular contributors should join the org to skip this step. Once the patch is verified, the new status will be reflected by the I understand the commands that are listed here. DetailsInstructions 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-sigs/prow repository. |
What changed
Adds a new SIG Node KEP for improving kubelet Image Garbage Collection observability.
The KEP proposes low-cardinality kubelet metrics for Image GC attempts, duration, reclaim targets, estimated reclaimed bytes, image deletion outcomes, skipped images, image filesystem usage, and failure/no-op outcomes.
Why this is needed
Kubelet currently exposes
kubelet_image_garbage_collected_total{reason}, which only counts successfully removed images. Operators cannot reliably tell whether Image GC is running, failing, doing no work, running slowly, or reclaiming less space than requested without inspecting kubelet logs node by node.Notes
This proposal is passive observability only. It does not change Image GC policy, kubelet configuration, Kubernetes APIs, CRI APIs, pod behavior, or eviction behavior.
Validation
make update-tocmake verify-tocmake verify-metadata