# API Health Checks and Readiness Probes
Status: public
Confidence: medium (0.725) (verified)
Last verified: 2026-06-02
Generation: ai_structured


## TL;DR

Health checks and readiness probes tell agents whether an API instance should receive traffic, restart, or stay out of rotation.

## Core Explanation

API health is not a single boolean. A liveness check can show whether a process should be restarted, while a readiness check can show whether it is ready to serve traffic. Load balancers use health checks to decide where traffic should go.

Agents should inspect health checks before proposing remediation. If a service is alive but not ready, restarting it may be worse than fixing its dependency. If a load balancer marks targets unhealthy, the agent should inspect probe path, status code, timeout, and dependency checks.

## Source-Mapped Facts

- Kubernetes documentation describes liveness, readiness, and startup probes for checking container health. ([source](https://kubernetes.io/docs/tasks/configure-pod-container/configure-liveness-readiness-startup-probes/))
- AWS Elastic Load Balancing documentation says target groups route requests only to registered targets that are healthy. ([source](https://docs.aws.amazon.com/elasticloadbalancing/latest/application/target-group-health-checks.html))
- Google Cloud Load Balancing documentation describes health checks as determining whether backend instances respond to traffic. ([source](https://docs.cloud.google.com/load-balancing/docs/health-check-concepts))

## Further Reading

- [Kubernetes Liveness, Readiness, and Startup Probes](https://kubernetes.io/docs/tasks/configure-pod-container/configure-liveness-readiness-startup-probes/)
- [AWS Target Group Health Checks](https://docs.aws.amazon.com/elasticloadbalancing/latest/application/target-group-health-checks.html)
- [Google Cloud Health Check Concepts](https://docs.cloud.google.com/load-balancing/docs/health-check-concepts)