Introduction

This is a demo of how to use KEDA and Kubernetes to scale RabbitMQ consumers based on the number of messages in the queue.

Steps

Enable k8s in docker desktop

Settings -> Kubernetes -> Enable Kubernetes

Run rabbitmq in local docker

```bash
docker run -d --hostname my-rabbit --name some-rabbit -p 5672:5672 -p 15672:15672 rabbitmq:3-management
```

Get rabbitmq container ip

```bash
docker inspect -f '{{range.NetworkSettings.Networks}}{{.IPAddress}}{{end}}' some-rabbit
```

write that ip in the consumer.py file
and the consumer-scaler.yaml file (this should go, for instance, in a config map in the real application)

Install keda in the k8s cluster

```bash
kubectl apply --server-side -f https://github.com/kedacore/keda/releases/download/v2.11.0/keda-2.11.0.yaml
```

Create the consumer docker image

In Visual Studio Code, right click on the Dockerfile file and select "Build Image"

Deploy the consumer to k8s

```bash
kubectl apply -f .\consumer-deployment.yaml
```

Deploy the scaler to k8s

```bash
kubectl apply -f .\consumer-scaler.yaml
```

Send messages to the queue

```bash
python .\producer.py
```

Check the scaler logs

```bash
kubectl describe hpa
```