I’ve been wanting to write this code for a while.  If I were as efficient at writing in Java, I would have done so, but, I’m slower and therefore wrote it in python.  It’s not very pythonic, but, whatever.


Metrobus is a framework that allows you to focus on your microservices application logic, not logic around pulling and pushing to Kafka.    Kafka for now.  Others to come soon.  All the details you could ever want on the github page.    I’ll be adding tickets in order to track some todo items.


Proof of concept and example for smart routing on a dumb bus.

This is a small project focused on my blog posts around routing on a message bus that’s dumb. Like Kafka.

Stateless and dumb: https://medium.com/capital-one-developers/stateless-and-dumb-microservices-on-a-message-bus-be78bca93ccb

Fast Cache: https://medium.com/capital-one-developers/blazing-fast-data-lookup-in-a-microservices-world-dd3ae548ca45

I use some of the caching ideas in here for a few of my data lookups.

The concept for the ‘test’ example application is fairly simple. The ‘pusher’ generates records, as if from a client or application. These records are simple JSON structures. For our example, we get cool things like an account number (FAKE!). The ‘pusher’ sends the message to the ‘Source’ topic on Kafka. Consider this your public entry point for upstream clients.


KNOWN TO DO ITEMS (Cause you know, SHIP IT!)

  • Simplify the logic or break up the code for the main handling function in metrobus.py
  • Determine best how to handle different situations like
    • Send to Error log
    • Send to dead letter queue
    • Dropped on purpose
  • Is it too simple?   what am I missing?