DNS Resolution Unreliable From Executor Pods

Problem

Executor pods intermittently fail to resolve admin-ui.ge-system.svc.cluster.local. Root cause unknown — may be CoreDNS cache or k3s single-node behavior. Workaround: host cron dumps k8s-health.json.

Impact

Discussion API calls fail intermittently. Discussion participation unreliable.

Proposed Solution

1. Check CoreDNS logs
2. Test with dnsutils pod
3. Consider dnsConfig in pod spec with explicit nameserver
4. Consider node-local-dns cache

Acceptance Criteria

• Executor pods reliably resolve k8s service DNS
• Remove host cron workaround