service-discovery

Service discovery replaces hardcoded URLs by maintaining a live registry of available instances with health status, allowing consumers to dynamically route requests to healthy services as they scale, deploy, or fail.

Health-aware service registration uses periodic liveness checks and heartbeat TTL pruning to automatically update instance health states—healthy, degraded, or unhealthy—and remove stale instances from the service registry.

Zero-downtime deployments use service discovery with self-registration, deregistration, and client-side load balancing to dynamically select healthy instances, ensuring reliable routing during blue/green or canary releases.

Multi-region routing applies to service discovery by storing region and capability metadata in the service registry, enabling consumers to select appropriate healthy instances based on weighted strategies across distributed architectures.

Client-side load balancing queries the service registry by service name and applies weighted selection strategies—prioritizing healthy instances and deprioritizing degraded ones—to choose an appropriate instance for routing requests.

A service registry is needed when microservice instances are ephemeral, requiring automatic failover, dynamic scaling, and health-aware routing that static configuration cannot reliably provide across resilient architectures.