AWS
339 incident problems in AWS environments.
먼저 읽을 가이드
추천 문제
All problems (339)
K8S-1554A cert-manager renewal starts and Route53 validation still failsDNS validation fails only for one delegated zone after renewal automation changes.KubernetesIntermediate10 minProK8S-1572A Karpenter node class updates and one nodegroup still launches with the old profileOne Karpenter nodegroup launches with the old instance profile after a rename.KubernetesIntermediate10 minProK8S-1528A Karpenter provisioner launches nodes and pods remain pendingPods stay pending on one scheduling path after a label schema migration.KubernetesIntermediate10 minProK8S-1518A Karpenter provisioner launches nodes and pods still stay pendingPods remain pending only on one scheduling path after a Karpenter label schema migration.KubernetesIntermediate10 minProK8S-1458An ExternalDNS rollout updates public records and one cluster path still flapsAn ExternalDNS rollout updates public records and one cluster path still flaps focuses on Service Discovery and asks the reader to isolate the key signal in AWS. Record flapping often comes from ownership-marker collisions rather than from...KubernetesIntermediate10 minProK8S-1487A cert-manager DNS01 challenge updates Route53 and still times outACME issuance fails intermittently only in hosted zones managed by both cert-manager and ExternalDNS.KubernetesIntermediate11 minProCICD-1528A Helm chart promotion succeeds and provenance verification still failsChart provenance checks fail only through one object storage proxy after promotion.CI/CDAdvanced11 minProCICD-1538A Helm chart promotion succeeds and provenance verification still failsProvenance checks fail only through one object-storage proxy after promotion.CI/CDAdvanced11 minProCICD-1548A Helm chart promotion succeeds and provenance verification still failsProvenance checks fail only through one storage proxy after promotion.CI/CDAdvanced11 minProCICD-1518A Helm release promotion succeeds and rollback failsRollback verification fails only after a chart repository object-versioning change.CI/CDAdvanced11 minProCICD-1461A reusable workflow keeps losing OIDC tokensA reusable deployment workflow works in its template repo and fails only when called from the protected production repo.CI/CDIntermediate11 minProCICD-1441A GitHub Actions deploy signs in to AWS and still failsA GitHub Actions deploy signs in to AWS and still fails focuses on Identity And Access and asks the reader to isolate the key signal in GitHub. OIDC failures around reusable workflows often come from environment-bound audience drift rather than...CI/CDIntermediate12 minProCICD-1421A GitHub Actions OIDC deploy works in one repo and fails in anotherA GitHub Actions OIDC deploy works in one repo and fails in another focuses on Identity And Access and asks the reader to isolate the key signal in GitHub. Reusable workflows can drift on requested audience values...CI/CDIntermediate12 minProCICD-1381A GitHub Actions OIDC deployment suddenly loses cloud authA deployment pipeline is reorganized by environment naming and later cloud auth breaks only for one promoted workflow.CI/CDIntermediate12 minProCICD-1499A Pulumi automation API run authenticates and state updates still failPulumi automation starts failing only after state bucket replication and object lock are enabled together.CI/CDAdvanced12 minProCICD-1509A Pulumi state backup succeeds and the restore test failsState restore drills fail only in the disaster recovery region after object lock policy changes.CI/CDAdvanced12 minProLINUX-1351A systemd timer writes backups fine by hand and fails overnightA systemd timer writes backups fine by hand and fails overnight focuses on execution-context and asks the reader to isolate the key signal in AWS. Scheduled jobs often fail because they inherit less context, not because the command itself is wrong.LinuxIntermediate12 minProLINUX-1349A timer-driven snapshot job works manually and misses AWS credentials at nightA timer-driven snapshot job works manually and misses AWS credentials at night focuses on execution-context and asks the reader to isolate the key signal in AWS. Successful manual CLI tests do not prove systemd timers inherit the same c...LinuxIntermediate12 minProCICD-1456A regional failover leaves CodePipeline manual approval stuckA regional failover leaves CodePipeline manual approval stuck focuses on Artifact Promotion and asks the reader to isolate the key signal in AWS. Approval failures after failover often hide in signed callback hostnames, not in the...CI/CDAdvanced13 minProCICD-1355A remote Terraform run fetches modules and fails provider authA remote Terraform run fetches modules and fails provider auth focuses on Identity And Access and asks the reader to isolate the key signal in AWS. Remote execution often fails on credential source precedence, not on missing cre...CI/CDIntermediate13 minPro