Service Registry

Service Registry is a library designed to simulate the 3rd parties that will invoke our distributed systems by isolating these external parties. It's eventually built upon a service registry pattern by implementing a basic circuit breaker

Context

This will give you a hint or context regarding what problems will be solved (or can be simulated) with this library

• How do the internal client will know about the available instances?
• How can we self-register one of the services down to a service still available?
• How do we handle the failure services without intersecting with other services?

Features

Given the context, we can implement several features that might help in the process of overcoming all of those problems such as:

• Circuit breaker
• Service registry (self-assigned services based on the available instance)
• Distributed services tracing (trace all of the failure and success counts)
• Built-in logging mechanism
• Gracefully shutdown a certain service

Usage

In the code there are several procedures to simulate this library, you can just run it by

$ python service_registry.py

And it will lead you to the logs output like :

15-09-2023 08:51:04 : service_registry.py : log : Success registered the service name
15-09-2023 08:51:04 : service_registry.py : log : Success registered the service name
15-09-2023 08:51:04 : service_registry.py : log : Success registered the service name
15-09-2023 08:51:04 : service_registry.py : log : Related service is healthy
15-09-2023 08:51:04 : service_registry.py : log : Related service is healthy
15-09-2023 08:51:04 : service_registry.py : log : Related service is healthy
{'GetSinglePost': {'url': 'https://jsonplaceholder.typicode.com/posts/1', 'assigned': False, 'assigned_service': None, 'availability': 'AVAILABLE'}, 'GetAllPosts': {'url': 'https://jsonplaceholder.typicode.com/posts', 'assigned': False, 'assigned_service': None, 'availability': 'STARTING'}}
15-09-2023 08:51:09 : service_registry.py : log : Related service is healthy
15-09-2023 08:51:09 : service_registry.py : log : Related service is healthy
15-09-2023 08:51:09 : service_registry.py : log : Related service is healthy
15-09-2023 08:51:09 : service_registry.py : log : Related service is healthy
{'total_requests': 1, 'successful_requests': 1, 'failure_requests': 0, 'duration': 0.0}
{'total_requests': 1, 'successful_requests': 1, 'failure_requests': 1, 'duration': 0.0}
15-09-2023 08:51:09 : service_registry.py : log : Success assigned one service to the available service
15-09-2023 08:51:09 : service_registry.py : log : Success assigned one service to the available service
15-09-2023 08:51:09 : service_registry.py : log : Success assigned one service to the available service
15-09-2023 08:51:09 : service_registry.py : log : Success assigned one service to the available service
15-09-2023 08:51:09 : service_registry.py : log : Success assigned one service to the available service
https://jsonplaceholder.typicode.com/posts
15-09-2023 08:51:09 : service_registry.py : log : Deleted GetSinglePost service instance
15-09-2023 08:51:09 : service_registry.py : log : Deleted GetSinglePost service instance
15-09-2023 08:51:09 : service_registry.py : log : Deleted GetSinglePost service instance
15-09-2023 08:51:09 : service_registry.py : log : Deleted GetSinglePost service instance
15-09-2023 08:51:09 : service_registry.py : log : Deleted GetSinglePost service instance
15-09-2023 08:51:09 : service_registry.py : log : Deleted GetSinglePost service instance

Or, you can adjust the procedure for running it in accordance with what you want to do (you might need to take a look in the built-in APIs directly from the source code to ensure what's the public method that you want to initialize)

Getting Started with the API

You need to install the Flask library to run the core API of this library, and the server starts on http://localhost:5000 by default. For example if you want to register a service, you can use the following API:

Register a service

curl -X POST http://localhost:5000/api/services \
  -H "Content-Type: application/json" \
  -d '{"service_name": "GetAllPosts", "service_url": "https://jsonplaceholder.typicode.com/posts"}'

Get an available service URL

curl http://localhost:5000/api/service/GetAllPosts

Simulate a service failure

curl -X POST http://localhost:5000/api/services/GetAllPosts/fail

View tracing metrics

curl http://localhost:5000/api/metrics

Caveats

• No async support. By all means, the service assignment is synchronous and resolves to the first available index. If that service is also unhealthy, no fallback will be attempted automatically
• In-memory state. All registry data is held in memory and does not persist across restarts

Acknowledgment

Perpetually, all of those implementations in this library are actually inspired by Hystrix-like from Netflix, and also the whole concept follows the instructions of the microservices pattern (you might read it here).

While it doesn't cover all of the crucial aspects that follow in the Hystrix, to be exact, this library describes what it would be like to simulate a service going down without causing other services to go as well