How to Build a Simple Service Discovery Registry in Python

A lightweight in-memory service registry class using a dict — register, deregister, and discover services with host, port, and version.

Easy Python 3.9+ Aug 9, 2026 System design patterns 13 views 0 copies

service-discovery registry dict microservices in-memory

Python code

36 lines

Python 3.9+

class ServiceRegistry:
    def __init__(self):
        self._services = {}

    def register(self, name, host, port, version="1.0"):
        self._services[name] = {
            "host": host,
            "port": port,
            "version": version
        }

    def deregister(self, name):
        return self._services.pop(name, None)

    def discover(self, name):
        service = self._services.get(name)
        if service:
            return f"{service['host']}:{service['port']} (v{service['version']})"
        return None

    def list_all(self):
        return self._services


if __name__ == "__main__":
    registry = ServiceRegistry()
    registry.register("auth", "10.0.0.1", 3000)
    registry.register("users", "10.0.0.2", 3001, version="2.3")
    registry.register("orders", "10.0.0.3", 3002)

    print(registry.discover("auth"))
    print(registry.discover("users"))
    print(registry.discover("missing"))
    print(registry.list_all())
    registry.deregister("orders")
    print(registry.list_all())

Output

stdout

10.0.0.1:3000 (v1.0)
10.0.0.2:3001 (v2.3)
None
{'auth': {'host': '10.0.0.1', 'port': 3000, 'version': '1.0'}, 'users': {'host': '10.0.0.2', 'port': 3001, 'version': '2.3'}, 'orders': {'host': '10.0.0.3', 'port': 3002, 'version': '1.0'}}
{'auth': {'host': '10.0.0.1', 'port': 3000, 'version': '1.0'}, 'users': {'host': '10.0.0.2', 'port': 3001, 'version': '2.3'}}

How it works

The registry stores each service as a dict keyed by its name, keeping host, port, and version together. register either adds a new service or overwrites an existing one, so a caller can update endpoints in place. discover uses .get() (not []), which avoids a KeyError when the service is missing and returns None instead. deregister uses pop(name, None) so it is idempotent — removing a non-existent service does not raise. The dictionary is returned directly from list_all, which is fine for a mock but should return a copy in real code to prevent callers mutating the registry.

Common mistakes

Using `self._services[name]` instead of `.pop(name, None)` in deregister, which raises KeyError for missing names
Returning the internal dict directly from `list_all`, letting callers mutate the registry by accident
Forgetting `.get()` in `discover` and raising KeyError when a service is not registered
Hard-coding version defaults without validating that the service data is complete

Variations

Use a `@dataclass` for the service entry instead of a plain dict
Add a `health_check` callback that probes each service endpoint before returning it

Real-world use cases

Stubbing an in-memory registry in unit tests so services can be registered and discovered without network calls.
Teaching the core pattern of service discovery before introducing ZooKeeper or Consul in a microservices codebase.
Starting a lightweight sidecar-style process that tracks internal microservice endpoints during local development.

How to Build a Simple Service Discovery Registry in Python

Python code

Output

How it works

Common mistakes

Variations

Real-world use cases

More from System design patterns

Tutorials

Quizzes

Python code

Output

How it works

Common mistakes

Variations

Real-world use cases

More from System design patterns

Keep learning

Tutorials

Quizzes