Scaling in terms of AKS simply means increasing or decreasing the number of pods (replicas) for each application instance. You can choose to manually scale the number of replicas up or down, or you can configure autoscaling based on utilization.
Before you scale your deployments, you should also ensure that you have sufficient capacity on the AKS nodes so that they can take that much load. If not, you should add new nodes first and then scale the application.
Manually scaling a deployment is very easy; you just need to modify the number of replicas in the YAML deployment file and submit the updated configuration using kubectl apply -f filename.yaml. This doesn't have an impact on any existing deployment and provisions the new pods based on the numbers defined.
Kubernetes services watch new pods that meet their label requirements...
