Why it matters
Development and test databases are often left running overnight and at weekends when nobody uses them.How nOps finds it
This recommendation comes from AWS Compute Optimizer through Cost Optimization Hub, not from an nOps rule. AWS flags an instance as idle from its connections and utilization over the lookback period (14 days by default). The saving is AWS’s estimate of the instance-hour charge you avoid while it is stopped.Before you start
Permission to stop and start the instance (rds:StopDBInstance, rds:StartDBInstance).
How to implement
- Confirm nobody needs the database while it is stopped.
- Stop it (RDS console → Databases → Actions → Stop temporarily, or
aws rds stop-db-instance --db-instance-identifier DB_ID). For Aurora, stop the cluster (aws rds stop-db-cluster). - For a database needed only during working hours, schedule the stop and start, for example with AWS Instance Scheduler or an EventBridge schedule.
- If nobody needs it at all, delete it with a final snapshot instead.
Risks and considerations
- RDS starts a stopped instance again after 7 days. Use a schedule, or delete the instance if it is not needed.
- Storage, provisioned IOPS and backups are still billed while it is stopped.
- To roll back, start the instance. It takes a few minutes to become available.
FAQs
Why did my instance start again on its own?
Why did my instance start again on its own?
RDS starts a stopped instance after 7 days so it can apply maintenance. Stop it again on a schedule, or delete it if it is not needed.