Challenge 25 · intermediate · Rollout and probes
Readiness vs liveness
A feed service runs as four pods behind a Service, serving 400 requests a second; each pod handles 150. At 10 seconds v2 rolls out. A new pod takes 5 seconds to start listening and 15 more to warm up its caches, and while warming it is eight times slower and can take a fifth of its normal load. The Deployment uses the Recreate strategy, has no readiness probe, and a liveness probe that restarts a pod whose /healthz takes longer than 1 second three times running.
From 10 to 120 seconds, answer at least 99% of requests with a p99 of 300 ms or less.
Goal · measured 00:10–02:00
0 changes · par 2- Requests answered ≥ 99.0%not measured yet
- p99 latency ≤ 300 msnot measured yet
Service map
What changed
Turn on a fault or change a defence while the run plays. Its effect shows here: the 5 seconds before against 5 seconds after it settles.
This is a simulation. Its numbers come from a model of a typical system, not from measuring real servers. The same link and seed always give the same run. How the Chaos Lab works.