When failover isn’t safe: Building high-availability PostgreSQL on Kubernetes
Summary
During a simulated "gameday" failure, Datadog discovered that network latency caused PostgreSQL replication lag to exceed safety thresholds, preventing the system from safely promoting standby nodes. To resolve this, they rearchitected their Kubernetes-based clusters to use synchronous replication for failover candidates, ensuring that automated failover can occur without risking data loss.
Read the Original Article
This article originally appeared on Datadog | The Monitor blog.
Read Full Article on Original Site
