Skip to content

CoreDNS Deployment: rolling update with replicas=1 and maxUnavailable=1 risks DNS downtime #85

Description

@coderabbitai

Summary

The CoreDNS Deployment in charts/kuadrant-instances/templates/coredns/01-deployment.yaml is currently configured with replicas: 1 and rollingUpdate.maxUnavailable: 1. This means that during a rolling update, the single DNS pod can be terminated before a replacement is running, causing a DNS service interruption.

Suggested Fix

Either:

  • Set maxUnavailable: 0 and add maxSurge: 1 to ensure the old pod is only removed after the new one is ready, or
  • Make replicas configurable so HA deployments can avoid this issue entirely.

Context

The current configuration intentionally mirrors the setup described in the upstream documentation. This issue is tracked here as a follow-up so the downtime risk is not lost.

References

Requested by @azgabur.

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

Labels

No labels
No labels

Type

No type

Projects

No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions