How to run a reliable RabbitMQ on K8S?

Abstract

You have a single RabbitMQ node running on Kubernetes (K8S). TGIR S01E05 covered the getting started part well. Deploying RabbitMQ to Kubernetes: What’s Involved? blog post added more detail.

With the RabbitMQ on K8S basics understood, it’s time to deploy a RabbitMQ cluster and tackle more advanced topics:

1. What are good liveness & readiness probes?
2. How to configure RabbitMQ for availability during RabbitMQ upgrades?
3. How to configure RabbitMQ for availability during K8S upgrades?
4. How to configure clients for handling a minority of RabbitMQ nodes becoming unavailable?
5. What to expect when a majority of RabbitMQ nodes go away?
6. What happens when all RabbitMQ nodes go away?