Problem
A process can be alive while its database, cache or required dependency is unavailable.
Engineering Context
Operational applications may depend on data services or external infrastructure. A basic health endpoint and a dependency-aware readiness endpoint answer different questions.
Design Considerations
Health should be cheap and stable. Readiness should reflect only dependencies that are genuinely required to serve normal traffic.
Architecture
A load balancer or operations check can use readiness while basic process monitoring uses health.
Trade-offs
Overly strict readiness can remove healthy instances because of a non-critical dependency; overly weak readiness can send traffic to an unusable application.
Implementation Pattern
Keep /health minimal. Use /ready for critical dependencies and return safe diagnostics without secrets.
When to Use It
Useful for backend services, APIs and applications with critical data dependencies.
Common Mistakes
Putting expensive business queries in health checks; exposing credentials or raw stack errors; treating every external API as a readiness blocker.
Conclusion
The distinction creates clearer operational signals and safer deployments.