Service Discovery
91 incident problems about Service Discovery. Start with the reviewed ones.
먼저 읽을 가이드
추천 문제
All problems (91)
K8S-1496A CoreDNS rollout succeeds and pods still resolve stale service IPsOnly pods behind one node-local DNS cache keep resolving a retired service IP after a CoreDNS rollout.KubernetesIntermediate9 minProK8S-1506A CoreDNS rollout succeeds and pods still resolve stale service IPsOnly pods behind one node-local cache keep resolving a retired service IP after a DNS rollout.KubernetesIntermediate9 minProK8S-1526A CoreDNS update rolls out and one namespace still resolves stale endpointsOnly one namespace sees stale endpoints after a CoreDNS update.KubernetesIntermediate9 minProK8S-1536A CoreDNS update rolls out and one namespace still resolves stale endpointsOnly one namespace sees stale endpoints after a CoreDNS update.KubernetesIntermediate9 minProK8S-1566A NodeLocal DNSCache restart succeeds and one namespace still gets NXDOMAINOne namespace sees NXDOMAIN only after search suffix changes.KubernetesIntermediate9 minProK8S-1556A NodeLocal DNSCache rollout completes and one namespace still resolves stale service IPsOne namespace sees intermittent traffic to an old service IP after DNS cache rollout.KubernetesIntermediate9 minProK8S-1494A cert-manager renewal updates the secret and ingress still serves the old certificateGateway API traffic keeps serving an old cert after a secret renewal that works fine on classic Ingress.KubernetesIntermediate10 minProK8S-1504A cert-manager renewal updates the secret and the gateway still serves the old certificateGateway API traffic serves the old cert after secret renewal while classic ingress updates fine.KubernetesIntermediate10 minProLINUX-1447A PXE environment stops leasing on one imagePXE boot works on older images and fails only after a kernel refresh on the host bridge.LinuxIntermediate11 minProK8S-1194kubectl exec works but the in-cluster service call still times outA public policy example was adapted to production. Debug access into the pod works, but normal service calls still time out because the policy does not allow the actual source path.KubernetesIntermediate20 minProK8S-1184Service has healthy pods but zero usable endpointsA public answer suggests checking selectors first. Selectors are correct, but readiness is the real gate keeping endpoints empty.KubernetesIntermediate20 minPro