Readiness Check

Readiness probe: 200 only when critical deps answer, else 503.

Nothing in a deployed environment probes this route, deliberately. The ALB target group and the ECS container check both point at /health (liveness): Postgres and the cache cluster are shared by every task, so one dependency blip fails this check fleet-wide at once, and ECS would then recycle tasks holding in-flight generations and per-task execution cache (see the stickiness comment on aws_lb_target_group.main, QUA-2560) while that dependency is already struggling. Those failures surface through the ALB 5xx and TargetResponseTime alarms and the RDS and cache alarms in modules/monitoring instead.

Its live caller is qua dev cloud, which blocks local startup on this route until the Postgres and Redis containers answer -- see devtools/src/dev/readiness.ts.