How much downtime can I expect for my RDS for SQL Server Single-AZ or Multi-AZ instance scheduled maintenance window?

Short description

AWS performs periodic maintenance to the physical host hardware, database version, or the operating system of your Amazon RDS for SQL Server instance. Before maintenance occurs, you receive an email notification. This email details the affected resources and the scheduled timeframe of the maintenance.

Estimated downtime differs for Single-AZ instances and Multi-AZ instances. The downtime that you experience might change based on the type of maintenance that's performed.

Resolution

Note: If you receive errors when you run AWS Command Line Interface (AWS CLI) commands, then see Troubleshoot AWS CLI errors. Also, make sure that you're using the most recent AWS CLI version.

You can view the AWS Health Dashboard to check whether scheduled maintenance involves downtime.

View past or pending maintenance actions

To review past scheduled maintenance for your RDS instances, see Viewing Amazon RDS events.

To view pending maintenance actions for your RDS instances with the AWS Management Console or the AWS CLI, complete the following steps:

1. Open the Amazon RDS console.
2. In the navigation pane, choose Databases.
3. Select the RDS for MySQL DB instance or DB instance that you want to view maintenance actions for.
4. Choose the Maintenance & backups tab to view the list of pending maintenance actions under the Pending maintenance section or run the following command:

   $ aws rds describe-pending-maintenance-actions --region example-region-name

Note: Replace example-region-name with your specified AWS Region name.

You can adjust the time of your maintenance window. For more information, see Adjusting the preferred DB instance maintenance window.

When you schedule your maintenance window

During a maintenance window, plan a controlled restart of any RDS SQL Server instance that handles heavy workloads. The controlled restart minimizes service disruption. Schedule the maintenance window during a low-activity period. This prevents an impact from the temporary performance degradation after a restart. Performance degrades when the restart terminates user connections, recreates the tempdb database, and clears the buffer and plan caches.

Before the scheduled maintenance window, verify that no long-running transactions are active and confirm that the transaction log is within normal size thresholds. These factors affect crash-recovery duration. If you use Always On Availability Groups with your Multi-AZ deployments, ensure all secondary replicas are healthy and fully synchronized. Healthy secondary replicas support a clean failover path. 

After the maintenance window ends and the service restarts, monitor the database recovery phases described on the Microsoft website. Log the progress of redo and undo, Availability Group (AG) synchronization status, and overall resource status to validate the instance's stability. This approach ensures the maintenance is applied safely, recovery times are predictable, and the system returns to normal operational levels with minimal impact.

Hardware maintenance downtime

Single-AZ instances: The instance is unavailable for several minutes while maintenance is performed. The exact time the maintenance takes varies based on the instance class type and database size.

Multi-AZ instances: Multi-AZ instances are unavailable for the time that it takes the primary instance to failover. This failover takes no more than 60 seconds.

Database engine upgrade downtime

Single-AZ instances: The instance is unavailable for at least 15 minutes while maintenance is performed. The exact time the maintenance takes varies based on the instance class type and database size.

Multi-AZ instances: In Multi-AZ instances, both the primary and standby instances are upgraded. Amazon RDS performs rolling upgrades. The outage lasts only for the duration of a failover after the secondary is upgraded and promoted to primary. The original primary then becomes the secondary and has the upgrade applied to it.

For more information see Considerations for SQL Server upgrades.

OS maintenance downtime

Single-AZ instances: The instance is unavailable for several minutes while maintenance is performed. The exact time the maintenance takes varies based on the instance class type and database size.

Multi-AZ instances: In Multi-AZ RDS instances, both the primary and standby instances are upgraded. RDS performs rolling upgrades. The outage last only for the duration of a failover after the secondary is upgraded and promoted to primary. The original primary then becomes the secondary and has the upgrade applied to it.

Related information

What do I need to know about the Amazon RDS maintenance window?