Amazon Aurora Global Database switchovers or failovers cause existing zero-ETL integrations to stop data ingestion because they are bound to a specific primary cluster ARN. This solution uses an event-driven architecture with Amazon EventBridge and AWS Lambda to automate the detection of primary cluster changes and the recreation of the integration. It also provides a workflow for updating CloudWatch replication health alarms and validating the new data synchronization target in Amazon Redshift.
Key points
Aurora zero-ETL integrations rely on unique primary cluster ARNs, requiring manual re-binding after global database failovers.
Automation using EventBridge rules can trigger Lambda functions to detect primary cluster transitions in real-time.
Re-establishing zero-ETL integration requires updating replication health alarms and creating new target databases in the local Redshift instance.
Operational recovery necessitates a two-part approach: automated integration reconciliation followed by manual validation and client rerouting.