ArchitectureA supervised JVM-class runtime — OLTP on seven engines, OLAP on three. AI-native, MCP-native, observable as plain SQL.Read the architecture
Está viendo la edición Perú. Está viendo la edición Colombia. You're viewing the Pakistan edition. Cambiar a la edición global →Cambiar a la edición global →Switch to the global edition →

Read replicas: one pool, read-only traffic routed transparently

A replica cluster attaches at server level. Every pool from that server can hand out a read-only connection routed to a replica with primary fallback, and a monitoring table shows each replica's health.

Reports and dashboards read far more than they write. Sending those reads to a replica is the cheapest scaling an estate can buy, if the application does not have to know about it. This release makes the routing a property of the server.

Routing

  • Attached at the server A replica cluster is declared on the database server, and every pool from that server inherits it.
  • Symmetric API A caller asks for a read-only connection and receives one on a replica, with fallback to the primary. The ordinary call is unchanged, so callers need no topology knowledge.
  • Strategies Round robin, least loaded and failover, with backoff from one second to sixty, tested on Informix and PostgreSQL containers.
  • Reports opt in A read-only flag on a report object redirects its execution to replicas.

Observed

  • A replica table in the monitoring catalog with a health percentage, a health state of healthy, degraded, primary fallback or offline, and the topology to its server.
  • Marked on pools and connections, so a connection row says whether it is a replica.

The connection pool rewrite of January gave the platform a pool worth routing through. This is the routing.