Designing a reliable unplanned failover

T1Covered by 1 source · first reported by Databricks / ClickHouse / DuckDB / Supabase / Neon / PlanetScale at 05 Nov, 00:00 UTC

How Neki detects a failed Postgres primary, and chooses a new one from the WAL that survived.

Sources

• T1Designing a reliable unplanned failoverDatabricks / ClickHouse / DuckDB / Supabase / Neon / PlanetScale05 Nov, 00:00