ApsaraDB RDS manages routine database O&M for you — including high availability configuration and security patches — but certain operations can briefly interrupt your service or require action on your end. This page explains what to expect and how to minimize impact.
Operational limits and usage notes
Specification change
When you change the specifications of an RDS instance, the system may migrate data to a new instance. Incremental data generated during migration continues to sync to the new instance. The switchover occurs at the time you specify.
If service interruptions or transient connections occur, you cannot perform most operations related to databases, accounts, and network settings.
Impact by edition
| Edition | Impact |
|---|---|
| RDS Basic Edition | Service unavailable for 30 minutes or longer. No secondary instance is provided as hot standby. Schedule specification changes during off-peak hours. |
| RDS editions other than RDS Basic Edition | Transient connection of approximately 30 seconds when the change takes effect. Configure your application to reconnect automatically, or schedule changes during off-peak hours. |
| RDS instances with cloud disks (non-Basic Edition) — storage capacity increase only | In most cases, no transient connections occur. |
| RDS Cluster Edition for ApsaraDB RDS for MySQL — cluster with only one secondary node | The read-only endpoint is inaccessible until the specification change takes effect. Schedule changes during off-peak hours to maintain business continuity. |
Version upgrade
Upgrading the major engine version or updating the minor engine version may cause service interruptions or transient connections. For example: updating the minor engine version of MySQL 8.0 from 20230324 to 20230630, or upgrading from MySQL 5.5 to MySQL 5.6.
If service interruptions or transient connections occur, you cannot perform most operations related to databases, accounts, and network settings.
Impact by edition
| Edition | Impact |
|---|---|
| RDS Basic Edition | Service unavailable for 30 minutes or longer. No secondary instance is provided as hot standby. Schedule upgrades during off-peak hours. |
| RDS editions other than RDS Basic Edition | Transient connection of approximately 30 seconds when the upgrade takes effect. Configure your application to reconnect automatically, or schedule upgrades during off-peak hours. |
| RDS Cluster Edition for ApsaraDB RDS for MySQL — cluster with only one secondary node | The read-only endpoint is inaccessible until the secondary node upgrade takes effect. Schedule upgrades during off-peak hours to maintain business continuity. |
Failover
RDS High-availability Edition and RDS Cluster Edition include a secondary instance or secondary node as hot standby. If the primary RDS instance or primary node fails unexpectedly, the system fails over to the secondary within 30 seconds.
If data is inconsistent between the primary and secondary at the time of failure, failover takes longer. During failover, a transient connection of approximately 30 seconds occurs. Configure your application to reconnect automatically to prevent workload interruption.
Network type change
Changing the network type between classic network and virtual private cloud (VPC) disconnects the application server from your RDS instance and assigns a new IP address to the instance. After the change completes, update the endpoint configuration in your application immediately.
Data restoration
Before restoring data, back up any important data on the instance to prevent data loss. Restore data to a temporary RDS instance or a cloned RDS instance first. After verifying the accuracy of the restored data, migrate it back to the original instance.
Storage capacity
If storage capacity is exhausted, the RDS instance is automatically locked and cannot provide services. Check storage usage regularly.
If storage usage is abnormally high, see What do I do if my ApsaraDB RDS for MySQL instance is locked because its storage capacity is exhausted?
Performance
Check whether the following resource allocations meet your workload requirements. Upgrade the instance if they do not.
| Resource | What to check |
|---|---|
| CPU cores | Whether processing capacity is sufficient for peak query volume |
| Memory | Whether buffer and cache requirements are covered |
| IOPS (input/output operations per second) | Whether disk throughput handles your read/write patterns |
| Storage capacity | Whether remaining space is sufficient with headroom for growth |
| Max connections | Whether the connection limit supports your application pool size |
Also review your database configuration on a regular basis: check for slow queries, identify SQL statements that need optimization, and remove unnecessary indexes or add missing ones.
Limits on database engines
For engine-specific quotas and limits, see the documentation for your engine: