Horizontal scaling

Horizontal scaling allows you to distribute your RayDB PostgreSQL workload across multiple nodes, improving performance and reliability by adding more instances rather than increasing resources on a single instance.

When to Scale Horizontally

Consider horizontal scaling if:

• Your read operations are increasing significantly.
• You need high availability with replication.
• Your application requires geo-distributed access with lower latency.

How to Scale Your Cluster

Read Replicas

1. Log in to the RayDB Dashboard.
2. Navigate to the Clusters section and select your primary cluster.
3. Go to the Replication & Scaling tab.
4. Click Add Read Replica and select a region if needed.
5. Confirm the setup to deploy additional read replicas.

Sharding (Partitioning Data)

For applications with high write workloads, consider sharding:

• Implement application-level sharding by distributing data across multiple clusters.
• Use PostgreSQL’s built-in table partitioning for managing large datasets efficiently.

Load Balancing

• Use a connection pooler such as PgBouncer to distribute read queries efficiently.
• Implement application-level logic to direct read queries to replicas and writes to the primary cluster.

Impact on Availability

• Read replicas are deployed asynchronously and provide non-blocking scaling.
• Application logic may need to be adjusted to handle distributed reads.
• Writes remain limited to the primary instance in a traditional setup.

Best Practices

• Monitor Query Performance – Use built-in monitoring to identify performance bottlenecks.
• Distribute Read Workloads – Direct read-heavy operations to replicas.
• Plan for Failover – Configure automatic failover strategies for high availability.

For more on optimizing performance, see Performance Tuning.