Support for HPA

[rms1000watt]

Would you be able to add support for Horizontal Pod Autoscalers? A workflow could be something like:

1. Check if HPA exists
2. Backup HPA (if exists)
3. Delete HPA
4. Do the normal canary deployment process
5. Deploy backed-up HPA with updated deployment version

Example deployment, service, HPA (not tested)

kind: Deployment
apiVersion: apps/v1
metadata:
  name: app-name-0.0.0
  labels:
    service: app-name
spec:
  selector:
    matchLabels:
      service: app-name
  replicas: 6
  template:
    metadata:
      labels:
        service: app-name
    spec:
      containers:
        - name: app-name
          image: 000000000000.dkr.ecr.us-east-1.amazonaws.com/app-name
          ports:
          - containerPort: 9000
---
kind: Service
apiVersion: v1
metadata:
  name: app-name
  labels:
    service: app-name
    version: 0.0.0
spec:
  ports:
  - name: http
    port: 9000
    protocol: TCP
    targetPort: 9000
  selector:
    service: app-name
---
kind: HorizontalPodAutoscaler
apiVersion: autoscaling/v2beta1
metadata:
  name: app-name
spec:
  scaleTargetRef:
    apiVersion: apps/v1beta1
    kind: Deployment
    name: app-name-0.0.0
  minReplicas: 6
  maxReplicas: 12
  metrics:
  - type: Resource
    resource:
      name: cpu
      targetAverageUtilization: 40

[kostis-codefresh]

Yes, we can add this. We are currently discussing internally if we are going to proceed with the current design of the canary step or if it is time to re-write it in Go/Python.

[rms1000watt]

https://github.com/codefresh-io/k8s-canary-deployment/pull/7 should be taken care of here

[kostis-codefresh]

Thank you.