Glossary · Automation software engineering and architecture
Readiness probe
Also known as: Readiness check
German: Bereitschaftsprüfung
In container orchestration, a readiness probe is a periodic check that determines whether a running service instance is ready to receive requests; instances that fail it are temporarily removed from traffic without being restarted.
- Software engineering
In one sentence
A readiness probe checks whether a service instance can accept requests; failing instances are taken out of traffic, not restarted.
Example
The data ingestion service reports not ready until its connection to the historian database is established, so the orchestrator sends no machine data to it yet.
How it applies
- Engineering: A readiness endpoint should check what the service needs to serve requests, such as loaded configuration and reachable dependencies, and respond quickly.
- Operation: Readiness probes enable rolling updates without interruption: new instances receive traffic only after they report ready.
- Documentation: Operations documentation should state what each probe checks and what a failing probe means, so on-call staff can distinguish a starting service from a broken dependency.
Readiness probe vs. liveness probe
A Liveness probe asks whether the instance is still working at all; failing it leads to a restart. A readiness probe asks whether it can serve right now; failing it only stops traffic. Mixing them up causes needless restarts.