Question

A company runs an SSM Automation runbook triggered by an EventBridge rule whenever an RDS instance emits a low-storage event. The runbook has three steps: (1) modify the instance to increase allocated storage, (2) wait for the instance to return to 'available', and (3) send an SNS notification. During a recent event, step 1 failed because the instance was already in a 'modifying' state from a prior automated run, and the entire automation aborted before sending any notification—leaving the on-call team unaware of the failure. The DevOps team wants the runbook to always notify the team of the outcome, whether the storage modification succeeds or fails, without abandoning the auto-remediation logic. What is the MOST appropriate way to achieve this?

Accepted Answer

Configure the SNS notification step as the runbook's onFailure target using the 'step:' syntax and also chain it as the normal next step after success, so it executes on both paths.

Answer

Add an onFailure attribute set to 'Continue' on step 1 and configure the notification step's isEnd to true so it always runs, but leave the wait step unchanged.

Answer

Increase the maxAttempts and set a timeoutSeconds on step 1 so the modification retries until the instance leaves the 'modifying' state, guaranteeing eventual success.

Answer

Enable EventBridge retry with a dead-letter queue on the rule so failed automations are reprocessed later and notifications are eventually delivered.