One PolarDB-X 1.0 is a distributed instance that consists of multiple compute nodes. Each node provides the features, such as Structured Query Language (SQL) routing, data merging, and data aggregation. You can change the node specification of the PolarDB-X 1.0 by changing instance configurations. You can upgrade the configurations during peak hours to share queries per second (QPS) of the business traffic. You can downgrade the configurations during off-peak hours to avoid wasting resources.
Notes
- Transient connection errors may occur in a short time when you downgrade the configurations due to the disconnection between the application and the instance. Make sure that your applications have the reconnection mechanism.
- When a persistent connection is in use, the newly added nodes after you upgrade the configurations cannot immediately receive traffic. We recommend that you restart Elastic Compute Service (ECS) instances in phased mode. Specifically, restart the ECS instances that host few services first, and then observe the service status to confirm that no problems occur before you restart all the other ECS instances.