Multigres private alpha: automatic failover for Supabase Postgres
Summary
- Supabase announced the private alpha of Multigres, bringing automated high availability and failover to Postgres projects.
- Multigres runs a three-node Postgres cluster across availability zones, managing routing, connection pooling, replication, and backups without modifying standard Postgres.
- Failover is handled as a consensus problem using synchronous replication to guarantee that acknowledged commits survive node failures without data loss.
- The built-in connection pooler automatically tracks session state per request, eliminates pooling mode selection, and includes fair-share resource allocation.
Why it matters
Multigres simplifies Postgres high availability and connection pooling on Supabase by automating multi-node consensus failover and removing the operational overhead of third-party tools.