RDS Recovery Mode

Question

Since a couple of weeks one of our production RDS machines keep rebooting. In the Events and Logs tab I see the following message:

"Recovery of the DB instance has started. Recovery time will vary with the amount of data to be recovered."

We have set Maintenance Window and Back-up Window to run after midnight, but these recoveries happen at different times during business hours.

What we need to know is;
1. How is this constantly triggered?
2. How can we prevent our production website to be down due to this recovery?

Iulian Constantinescu · Answer

RDS recovery can be triggered for several reasons:

1. Hardware failure: The underlying hardware hosting your RDS instance may have experienced a failure
2. Storage issues: Problems with the storage volume attached to your instance
3. Network connectivity issues: Network disruptions between components
4. Database engine crashes: The database process itself crashed due to bugs or resource constraints
5. Monitoring system detection: AWS's automated monitoring detected an unhealthy state

Investigate Root Cause
To address the recurring reboots:
* Check CloudWatch metrics for unusual patterns (CPU, memory, storage, connections)
* Review RDS logs for errors before the recovery events
* Check if you're hitting resource limits (connections, storage)
* Consider upgrading your instance type if it's resource-constrained

The most effective solution is to enable Multi-AZ for your RDS instance. This creates a standby replica in a different Availability Zone that automatically takes over during failures:

* During planned maintenance or instance failures, RDS automatically fails over to the standby
* Failover typically completes within 60-120 seconds
* Your application connects to the same endpoint, so no code changes are needed

If the issue persists after implementing Multi-AZ, contact AWS Support for deeper investigation.

RDS Recovery Mode

Add your answer

Relevant content