13  Scaling

Scaling EdgeSet is different than scaling a traditional data warehouse. The size of the EdgeSet server must scale with the number of concurrent queries it can run, rather than the amount of data it has access to.

13.1 Read replicas

The primary method of scaling EdgeSet for most data sources is through read replicas. Since EdgeSet offloads queries whenever possible to the data source, the performance of the EdgeSet-facing database will determine query performance.

A simple architecture for high scalability and minimal disruption to other applications is to create a dedicated EdgeSet read replica for all data sources that support them.