Configure startup containers for Intel deployment

[danc094codetogether]

Description

A customer is requesting to add the ability to configure startup containers for the Intel deployment within the helm chart.

Having the ability to add startup containers would allow customer to make the Intel Pod to wait for Cassandra to be ready before trying to start the application (which would make deployments more seamless).

The customer has indicated this is not a high priority but a very nice to have.

[danc094codetogether]

Default behavior (initContainers not used)

• Goal: Confirm backward compatibility when initContainers is empty (default).

Deploy Intel without overriding initContainers (leave it as default []):

helm upgrade --install intel ./chart-path -f values.yaml

Verify the pod has no initContainers:

kubectl get pods
kubectl describe pod -l app.kubernetes.io/instance=intel | sed -n '/Init Containers:/,/Containers:/p'

• Expected: Init Containers: is absent or empty.
• Verify rollout succeeds and pod becomes Ready:

kubectl rollout status deploy/intel
kubectl get pods

• Expected: Pod is Running and Ready.

Wait for Cassandra (using the values.yaml example)

• Goal: Confirm the initContainer blocks Intel startup until Cassandra is reachable.
• Create an override file qa-values.yaml with the same example from values.yaml:

# Useful to wait for Cassandra or run startup checks before Intel starts.
initContainers:
  - name: wait-for-cassandra
    image: busybox:1.36
    command: ["sh", "-c"]
    args:
      - >
        until nc -z codetogether-cassandra.default.svc.cluster.local 9042;
        do echo "Waiting for Cassandra..."; sleep 2; done

• Install/upgrade Intel with the override:

helm upgrade --install intel ./chart-path -f qa-values.yaml

• Verify the initContainer is present:

kubectl describe pod -l app.kubernetes.io/instance=intel | sed -n '/Init Containers:/,/Containers:/p'

• Expected: You see wait-for-cassandra under Init Containers.
• Observe pod behavior while Cassandra is NOT reachable:

kubectl get pods -w

• Expected: Pod stays in an init phase (e.g., Init:0/1) and the main Intel container does not start.
• Check initContainer logs:

POD=$(kubectl get pod  -l app.kubernetes.io/instance=intel -o jsonpath='{.items[0].metadata.name}')
kubectl logs "$POD" -c wait-for-cassandra

• Expected: Repeated lines like Waiting for Cassandra...
• Make Cassandra reachable (use whatever method matches the QA environment):
• Start/scale Cassandra up, or
  Fix service/DNS/port so codetogether-cassandra.default.svc.cluster.local:9042 becomes reachable.
• Verify Intel starts after Cassandra is reachable:

kubectl rollout status deploy/intel
kubectl get pods

• Expected: InitContainer completes, Intel container starts, pod becomes Running and Ready.

Negative path (Cassandra never becomes reachable)

• Goal: Confirm Intel does not start if the initContainer cannot complete.
• Keep Cassandra down/unreachable (or deploy in an environment where that DNS/Service does not exist).
• Deploy using the same qa-values.yaml from Test Case 2.
• Verify:

kubectl get pods
kubectl describe pod -l app.kubernetes.io/instance=intel
kubectl logs "$POD" -c wait-for-cassandra

• Expected: Pod remains in init state indefinitely, with logs continuing to print Waiting for Cassandra....

@DavidRS13

[DavidRS13]

Tested ✅