kube-node-ready-controller: Update to version master-68

[k8s-on-aws-manager-app]

• https://github.com/zalando-build/kube-node-ready-controller/pull/84

[zalando-robot]

🤖 LLM Suggestions

[High] PR description lacks context on what master-68 changes

The PR description only links to an upstream PR (zalando-build/kube-node-ready-controller#84) without summarising what that PR introduces (bug fix, feature, security patch, etc.). Reviewers and on-call engineers cannot assess the risk of this rollout without knowing what changed in the new image. Please add a short summary of the upstream change and its expected production impact to the PR description.

Other non-critical findings

[Low] No explicit readinessProbe on the DaemonSet pod

The manifest does not define a readinessProbe for the controller container. For a DaemonSet that runs on every node and exposes a /metrics endpoint on port 9090, a minimal HTTP readiness probe would ensure traffic (and node-readiness decisions) are only taken after the controller has fully started. Consider adding:

readinessProbe:
  httpGet:
    path: /healthz
    port: 9090
  initialDelaySeconds: 5
  periodSeconds: 10

(Adjust the path to match whatever health endpoint the controller exposes.)

Tokens Used: Input: 3,988 | Output: 376 | Cache Write: 3,986 | Cache Read: 0 | Total cost: $0.0206

[mikkeloscar]

👍

[katyanna]

👍