AWS
339 incident problems in AWS environments.
먼저 읽을 가이드
추천 문제
All problems (339)
CICD-1375A remote Terraform run fetches modules and still fails provider authA remote Terraform run fetches modules and still 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 rather than on mi...CI/CDIntermediate13 minProCICD-1438A CodeBuild artifact publish succeeds and downstream deploy failsA CodeBuild artifact publish succeeds and downstream deploy fails focuses on Artifact Promotion and asks the reader to isolate the key signal in AWS. Storage lifecycle changes can alter metadata assumptions downstream d...CI/CDAdvanced14 minProCICD-1448A CodeBuild artifact reaches S3 and deployment failsA CodeBuild artifact reaches S3 and deployment fails focuses on Artifact Promotion and asks the reader to isolate the key signal in AWS. Artifact storage optimizations can quietly change metadata contracts that deployment steps dep...CI/CDAdvanced14 minProCICD-1428A CodeBuild multi-stage pipeline passes build and fails deploymentA CodeBuild multi-stage pipeline passes build and fails deployment focuses on Artifact Promotion and asks the reader to isolate the key signal in AWS. Artifact retention optimizations can subtly alter metadata contracts...CI/CDAdvanced14 minProCICD-1291A generated workflow fails only in one repoA repo adopts a shared deployment template and only later discovers cloud auth fails in the copied version.CI/CDIntermediate14 minProCICD-1347A GitHub Actions OIDC exchange fails only on scheduled runsAn organization standardizes deployment workflows and later discovers cloud auth fails only for schedules while push-based deploys continue working.CI/CDIntermediate14 minProK8S-1298A CronJob overlaps unexpectedlyA scheduled batch task starts producing duplicate side effects during node churn or maintenance windows.KubernetesIntermediate15 minProK8S-1282A DaemonSet never reaches the last nodeAn agent DaemonSet covers nearly the whole cluster and one newly introduced node group never picks it up.KubernetesIntermediate15 minProK8S-1290A Job runs successfully once and then flakesA batch process or long-running job reads a mounted secret once and later begins failing after an external rotation event.KubernetesIntermediate15 minProK8S-1296A node becomes Ready but DNS for hostNetwork pods failsA node image update appears safe and later hostNetwork workloads alone begin failing name resolution.KubernetesIntermediate15 minProK8S-1278A pod can reach services by IP but not by nameA monitoring or system pod using hostNetwork starts failing service discovery even though raw IP connectivity is fine.KubernetesIntermediate15 minProK8S-1286A readiness probe passes manually but fails in clusterA pod with a service-mesh or sidecar proxy starts answering locally and still fails readiness for the first minutes after launch.KubernetesIntermediate15 minProCICD-1276A release job creates the image correctly but the deploy host still pulls the previous versionA release pipeline reports success and later the runtime still runs yesterday's image even though the registry contains the new one.CI/CDIntermediate15 minProK8S-1311A StatefulSet rollout stalls only on replacement nodesA storage add-on update seems fine until stateful workloads land on new nodes and never attach volumes successfully.KubernetesIntermediate15 minProK8S-1292A validating webhook starts timing out only after a NetworkPolicy rolloutA security hardening rollout adds NetworkPolicies and later new workloads cannot be created due to webhook timeouts.KubernetesIntermediate15 minProK8S-1284An ALB ingress rule looks correct but one path returns 404A working ingress is migrated to a newer API version and only one nested route stops matching even though the rule appears equivalent.KubernetesIntermediate15 minProK8S-1276An ingress provisions but health checks stay redA service behind an ALB opens in the browser but keeps rotating targets in and out because health never goes green.KubernetesIntermediate15 minProCICD-1284A deployment job can push images but cannot update the registry cacheA team adds registry-backed Buildx caching and only the cache step begins failing while the ordinary push still works.CI/CDIntermediate16 minProCICD-369A preview environment uses feature flags from production defaults and hides one missing secret until the production rollout toggles the path live during a staged decommissionLower environments looked clean only because the risky path never became active there. The service still works through the primary path, but one dependency only fails when the old component is finally drained away.CI/CDIntermediate16 minProK8S-1280A Service works inside the cluster but NLB traffic failsA LoadBalancer Service behaves perfectly in-cluster and starts failing only for outside clients after NLB source preservation enters the path.KubernetesIntermediate16 minPro