What is service discovery and why do microservices need it?

Service discovery is the mechanism that lets one microservice find the current network endpoints of another by a stable logical name (e.g., payments-api) instead of a hard-coded IP or port.

Why microservices need it:

• Pod / container IPs are not stable. Every deploy, restart, autoscale, or node failure can change them.
• Hard-coded URLs break the first time something moves.
• You need a level of indirection: caller asks for a name → registry returns one or more live endpoints.

What it solves (four things at once):

1. Naming — services find each other by stable name, not IP.
2. Health — only healthy instances are returned.
3. Load balancing — calls are distributed across instances.
4. Failure handling — dead instances drop out of the registry within seconds.

Two patterns you''ll meet:

• Client-side (Netflix Eureka, Consul) — caller queries the registry and picks an instance itself.
• Server-side (Kubernetes Services, AWS ELB) — caller hits a stable LB; the LB does the lookup.

In 2026, the default for new microservices is server-side via Kubernetes Services + cluster DNS — no extra registry needed.