Routage des microservices selon le chemin

In a microservice system, clients should hit a single host while Nginx fans requests out to the right service based on the URL path.

This keeps clients simple and hides the internal topology.

Routing decisions live in location blocks. Each matches a path pattern and proxies to a specific upstream.

Name each service in an upstream block so locations stay readable and you can scale instances later.

A plain location /path/ is a prefix match. Use = for an exact match, which is faster and avoids accidental overlaps.

A trailing slash on proxy_pass changes path rewriting. With proxy_pass http://svc/; the matched prefix is stripped; without it the full path is forwarded.

Prefix with ~ for case-sensitive regex matching, useful for routing by file type or dynamic segments.

Nginx evaluates locations in a defined order: exact =, then literal prefixes, then regex in file order. Understanding this prevents surprising routes.

Add a fallback location to serve a friendly response for unknown paths instead of leaking a default error.

Keep forwarding headers consistent across all services by including a shared snippet in each location.

Onboarding a service is just a new upstream plus a new location, no client changes required, which is the whole point of a gateway.

After reloading, curl each path and confirm it reaches the intended service.

With location /users/ { proxy_pass http://users-service/; }, what path does the upstream receive for a request to /users/42?

You built path-based routing to multiple services:

• location blocks map paths to upstreams
• Trailing slash on proxy_pass controls prefix stripping
• Match types: exact =, prefix, and ~ regex
• A catch-all handles unknown paths

One host now cleanly fronts an entire fleet of microservices.