Problem
Paradox Shop's checkout writes flow through a connection pool to a single PostgreSQL primary at 1,000 RPS. That primary needs to be able to fail, whether from a bad deploy, a host dying, or anything else, without turning into a multi-hour outage or ambiguity about which node is actually in charge afterward. Build the pooled write path, and designate exactly one replica for automatic, data-safe failover.
Required components
Click to add a component to the canvas.
Additional components
Browse by category. Click or drag a component onto the canvas.