Alive is not the same as ready for traffic#
A liveness signal asks whether restarting the process might help. Readiness asks whether this instance should currently receive traffic. The vocabulary is broadly useful even outside a particular orchestrator.
| Too shallow | Too deep |
|---|---|
| Returns 200 while a critical local resource is unusable | Calls every downstream service and couples all health |
| Routes traffic to a broken instance | One database hiccup can evict every API instance |
Keep the check cheap, bounded, and aligned with the action its consumer will take. A restart probe and a load-balancer routing probe may need different evidence.
Related practice problems
Use the manifest-backed prompts below to apply this lesson in a complete design.
- Health Endpoint Monitoring patternMicrosoft Azure Architecture Center
- Addressing cascading failuresGoogle SRE
Reach this point to mark the lesson as read.