Per User Counter but using microservices

Architecture:

• Gateway:
  • Exposes HTTP endpoints
  • Checks for and verifies JWT token
  • Communicates with Auth and Counter using gRPC
• Auth:
  • Handles Login and Registration
  • Uses PostgreSQL as database and gRPC to communicate with Gateway
• Counter:
  • The main service
  • Communicates with Gateway using gRPC
  • Uses PostgreSQL as database and Memcached as a caching layer
• Logger:
  • Sends logs to an Elasticsearch instance running along with Kibana for visualisation
  • Collects logs from a RabbitMQ instance before sending to Elasticsearch
  • Uses a multithreading-safe worker pool to send several requests concurrently
• Logger Client:
  • Client library that provides an API for different services to send logs to the Logger service
  • Passes on the Log messages to the RabbitMQ instance to be picked up by the Logger

Logger can easily be scaled horizontally by multiplying the instances. Other services might require a bit more work.