Scalability
AnalyticDB for PostgreSQL supports both horizontal and vertical scaling to match your workload as it grows or fluctuates.
Scale-out
Two scale-out methods are available: scale-out within an instance and scale-out across instances.
Scale-out within an instance adds compute nodes to an existing instance. Data migration is minimized, so this method completes faster.
Scale-out across instances migrates data from an existing instance to a new one. Use this method when you need to scale across zones or regions. This approach offers greater placement flexibility but involves higher data migration costs.
Scale out coordinator nodes
To handle high-concurrency read and write workloads, scale out coordinator nodes. This enhances read and write capabilities in scenarios that involve high concurrency.
Scale-up or scale-down
Scale-up or scale-down adjusts the resources of individual compute nodes—disk capacity, CPUs, or memory—without migrating data. Configuration changes take effect within minutes, making this method well-suited for traffic bursts or temporary capacity needs.
Combined with elastic node scheduling, scale-up or scale-down helps maximize computing power utilization and accelerate query performance.
Choose a scaling method
| Scenario | Recommended method |
|---|---|
| Data volume is growing and you need more storage or processing capacity | Scale-out within an instance |
| You need to distribute capacity across zones or regions | Scale-out across instances |
| Read and write throughput is limited by high concurrency | Scale out coordinator nodes |
| Traffic spikes require quick capacity adjustments without data movement | Scale-up or scale-down |