AWS RDS instance stuck in rebooting state, no recovery

An RDS instance stuck in rebooting state for over 3 hours is abnormal and requires immediate attention. Here's what you need to know:

What could cause this issue:

Rebooting a DB instance normally results in only a momentary outage while the database engine service restarts. Extended rebooting states can be caused by:

• Underlying infrastructure issues
• Database engine problems preventing proper startup
• Resource contention or system-level issues
• Corrupted database state that's preventing the instance from coming back online

Troubleshooting steps:

1. Check RDS Events: Review the Amazon RDS events in the console at the time of the reboot and during the stuck period. Look for any error messages or warnings that might indicate the root cause.

2. Review CloudTrail logs: Use AWS CloudTrail to examine the API call history and event details around the time of the reboot to identify any anomalies.

3. Check database error logs: If accessible, review the database error logs for any error messages or warnings related to the restart failure.

What you can do:

Unfortunately, when an RDS instance is stuck in the rebooting state, your options from the customer side are limited. You cannot force a recovery or perform another reboot while the instance is in this state.

Recommended action:

Given the extended duration (3+ hours) and critical impact to your application, you should contact AWS Support immediately. This is not a normal situation, and AWS Support will need to investigate the underlying cause and potentially intervene at the infrastructure level to recover your instance.

While waiting for Support, you should prepare for potential recovery options:

• Identify your most recent automated backup or manual snapshot
• Be ready to restore from a snapshot if the instance cannot be recovered
• Document the exact timeline and any events you observed

Do not wait indefinitely - the longer the instance remains in this state, the more critical it becomes to have AWS Support involved to either recover the instance or advise on the best recovery path forward.
Sources
RebootDBInstance - Amazon Relational Database Service
Troubleshoot restarts and failovers in Amazon Aurora | AWS re:Post