//This repo is still WIP// KEDA is a Kubernetes-based event-driven autoscaler. KEDA determines how any container in Kubernetes should be scaled based on the number of events that need to be processed. More information on KEDA here - https://keda.sh/
A video tutorial explaining the same is here - https://www.youtube.com/watch?v=-jOVplEzypY
The current project is a simple PoC to understand and demo KEDA in action.

The project has three directories as follows:
a. client: This directory has a file by the name kedastorclient.py. This file has the code to enqueue a pre defined number of messages to the Azure Storage Queue. The Azure Storage connection string, storage queue name and the number of messages to be enqueued need to be set in this file.

b. receiver: This directory has a file by the name kedastorreceiver.py. This file has the code to dequeue one message from the Azure Storage Queue. There is an indefinite while loop in the code just to stop the code from exiting after dequeuing the message. The Azure Storage connection string need to be set in this file along with the queue name.
The directory also has the Dockerfile to deploy the code as a container.

Source code reference for kedastorclient.py and kedastorreceiver.py taken from here - https://docs.microsoft.com/en-us/azure/storage/queues/storage-quickstart-queues-python
c. yaml: This directory has the deployment yaml file by the name kedastorreceiver-deployment.yaml for the container created from kedastorreceiver.py. In this deployment file, the name of the image created from kedastorreceiver.py and the Azure Storage connection string should be entered.
Another file in this directory is by the name scaledobject.yaml. More information here - https://keda.sh/docs/1.5/scalers/azure-storage-queue/. Care should be taken on the following points:

1. The deploymentName property should be set to the deployment name used in kedastorreceiver-deployment.yaml
   
2. The connection property should be set to the name of the key which is defined as an environment variable in kedastorreceiver-deployment.yaml.
   
3. The namespace of both -- the scaledobject as well as the deployment should be the same.
   

The following services need to be provisioned in Azure:

1. Azure Storage Queue: I have used the name of the queue as kedatest.
   
2. Azure Container Registry: To store the Docker images for the container created from kedastorreceiver.py.
   
3. AKS: AKS should have KEDA deployed. Steps for the same can be found at: https://www.thinktecture.com/en/kubernetes/serverless-workloads-with-keda/serverless-workloads-with-keda/
   . The container created from kedastorreceiver.py will be deployed here.
   

Executing kedastorclient.py will enqueue predefined number of messages to the Azure Service Bus Queue. The Scaled Object defined using scaledobject.yaml will scale the pods created of kedastorreceiver.py in the AKS based in the messages enqueued in the Azure Service Bus queue.