
Catch a deadlocked container that keeps its port open by giving it an HTTP liveness probe with tight thresholds, and prove the restart clears the wedge while a healthy service is left alone.
Read the full problemGate a cold cache out of rotation with a readiness probe, and prove the Ready condition tracks the warm-up at startup, through a mid-life failure and on recovery without a single restart.
Read the full problemProtect a slow boot with a startup probe carrying its own budget so a tight liveness probe can take over afterwards, and prove both the over-budget kill and the post-boot wedge land on their own timing.
Read the full problemProbe a fleet of four misbehaving services so each self-reports and self-heals honestly - tight liveness for the wedger, warm-up readiness for the cache, a startup budget in front of liveness for the slow booters - and audit conditions, restarts and probe failures across the namespace.
Read the full problem