When connections reach the upper limit or performance issues occur, you can restart the instance. The restart causes a ~30-second interruption, so prepare your application before proceeding.
When to restart
Restart your instance in the following situations:
-
Connection limit reached: The number of active connections hits the upper limit, blocking new connections.
-
Performance degradation: The instance becomes unresponsive or experiences significant slowdowns.
-
Stuck temporary files: On PostgreSQL versions earlier than PostgreSQL 14, a system failure may leave temporary files that cannot be deleted. A restart clears them.
Prerequisites
Before you begin, ensure that:
-
The instance is in the Running state
Impacts
Review the following impacts before restarting.
Transient connection interruption
The restart causes a ~30-second connection interruption. Make sure your application is configured to automatically reconnect. To minimize disruption, stop writes to the instance before restarting.
Minor engine version update
If the instance uses Premium Local SSDs, the system automatically updates the instance to the latest minor engine version after the restart.
Read-only instances
Read-only instances attached to the primary instance are not restarted. Data replication between the primary and read-only instances resumes automatically after the primary restarts.
RDS Basic Edition
RDS Basic Edition has no secondary instance as a hot standby. If your RDS instance runs RDS Basic Edition and it unexpectedly fails, or you restart the instance, change its specifications, or upgrade its database engine, your database service may become unavailable for an extended period of time. For workloads that require high availability, use RDS High-availability Edition instead. You can also upgrade from RDS Basic Edition to RDS High-availability Edition for instances of specific database engine versions. For details, see Upgrade the edition of an RDS instance from RDS Basic Edition to RDS High-availability Edition.
RDS High-availability Edition
Restarting the primary instance may trigger a primary/secondary switchover. After the restart, the zones of the primary and secondary instances may be switched.
Reduce restart time
Restart duration depends on the number of tables, databases, and active connections, and the current service traffic. To shorten the restart:
-
Close unnecessary active connections before the restart.
-
Reduce service traffic to the instance.
-
Avoid creating unnecessary tables or databases.
After the instance restarts, reconnect your services.
Restart an instance
After you click Restart Instance, the instance status changes to Restarting and workloads are interrupted. Complete all preparation steps before proceeding.
-
Go to the Instances page. In the top navigation bar, select the region where the instance resides. Find the instance and click its ID.
-
In the upper-right corner of the page, click Restart Instance.
-
In the confirmation dialog, click OK.
FAQ
Do endpoints or virtual IP addresses change after a restart?
No. Endpoints and virtual IP addresses remain unchanged after a restart.
Does restarting the primary instance restart read-only instances?
No. Read-only instances are not restarted. Data replication resumes automatically once the primary is back online.
If I use RDS High-availability Edition, does restarting trigger a primary/secondary switchover?
A switchover may be triggered. After the restart, the zones of the primary and secondary instances may be switched.
Are there situations other than connection limits and performance issues where I should restart?
Yes. On PostgreSQL versions earlier than PostgreSQL 14, a system failure can leave temporary files that the system cannot delete. Restarting the instance resolves this. If other issues occur on your RDS instance, you can also restart the instance to resolve them.
API reference
| Operation | Description |
|---|---|
| Restart an instance | Restarts an RDS instance |