The headline needs to capture: identical Kubernetes health probes, 10-minute outage became 40 minutes, the cause (misconfiguration).

Short summary

Misconfiguring Kubernetes health probes by pointing both liveness and readiness to the same endpoint triggered a 40-minute restart storm from a 10-minute database outage. The fix is separating health endpoints using Spring Boot Actuator groups: /liveness for process health only, /readiness for dependencies. Prevention requires process discipline—decide each new dependency's probe group in code review to prevent silent configuration drift.

• •Identical liveness and readiness probes cause Kubernetes to kill healthy pods when dependencies fail, extending outages via cascading restarts
• •Spring Boot Actuator health groups split concerns: /liveness answers 'is the process wedged' (process-internal only), /readiness answers 'can this pod serve traffic' (includes dependencies)
• •Configuration drift is the recurring threat—enforce probe-group decisions in code review when adding new dependencies to prevent the same bug resurfacing with new triggers

Generated with AI, which can make mistakes.