Why did my Amazon RDS DB instance fail over?

Short Description

If there is a planned or an unplanned outage for a Multi-AZ DB instance, Amazon RDS automatically switches to a standby replica or secondary instance in another Availability Zone. Depending on your database activity at the time of the time of the outage, failover usually lasts between 60-120 seconds. For more information, see Failover Process for Amazon RDS.

Resolution

To see if a failover has occurred, open the Amazon RDS console, and then choose Events from the navigation pane. If AWS CloudTrail logging is enabled, then you can check the logs to see whether the event was planned or unplanned. For example, scaling compute or applying pending OS upgrades can trigger a failover.

If there are no events listed and there are no CloudTrail logs, the failover might be caused by one of the following:

• Availability Zone failures
• Primary DB instance compute node failures
• Networking issues with the primary DB instance
• Storage or Amazon Elastic Block Store (Amazon EBS) volume issues

When these types of unplanned outages occur, Amazon RDS monitors and identifies the issue. Then, Amazon RDS fails over automatically to minimize downtime.

Related Information

Best Practices for Amazon RDS