Scenario answers

EventBridge action repeats

Assume at-least-once behavior and make the target idempotent. Validate current state before acting and preserve failed events in a DLQ.

Configuration drift event

Use AWS Config for compliance evaluation and Systems Manager Automation for a safe, scoped remediation. Define exceptions to prevent remediation loops.

Pipeline failure

Start with the failed stage and action execution details. Inspect the native service logs, artifact location, execution role, input/output configuration, and the earliest failed event.

CloudFormation rollback failure

Find the first failed resource event, correct the external cause, and use supported rollback recovery rather than making broad manual resource changes.

ECS tasks repeatedly stop

Inspect stopped reason, container exit code, service events, image-pull access, task role, execution role, networking, health checks, and capacity placement.

EKS pod is Pending

Inspect pod events, scheduling constraints, resource requests, taints, node capacity, quotas, persistent volumes, and admission/controller behavior.

Incident Manager in a new account

Do not assume it can be newly enabled. Since November 7, 2025, Incident Manager is closed to new customers. Use a replaceable incident layer based on generally available event, automation, OpsCenter, notification, collaboration, and third-party capabilities.