Mark ephemeral pods created as part of kanister functions with JobID label

[mabhi]

Change Overview

Due to an ongoing work regarding identifying ephemeral pods those get spawned during various operations as part of some of the kanister functions, we need to mark these pod with some labels. One such pre-decided label for these pods is kanister.io/JobID. This label will help:

• Developers to verfiy whether required pod-specs were applied to these pods.
• Identify pods, due to erroroneous cases stay lingering in memory without ever completing.
• If multiple such type of pods are instantiated and we want to target / debug one out of these.

Right now the flow of this label information is, with the instantiation of actionset against a blueprint that services various kanister functions is triggered from an external application, and let's say if this label populated with a JobID, then this label [key: value] pair is added in entirerity to the temporary pod, so we can track it.

This PR handle cases when the kube-task labelled pods invoke kanister function that has a context.Context instance passed that contains a key-value pair corresponding to the pre-decided label, then it will make its way to the pods label selector map in addition to others if present already.

Pull request type

Please check the type of change your PR introduces:

• 🚧 Work in Progress
• 🌈 Refactoring (no functional changes, no api changes)
• 🐹 Trivial/Minor
• 🐛 Bugfix
• 🌻 Feature
• 🗺️ Documentation
• 🤖 Test

Issues

• fixes #issue-number

Test Plan

• 💪 Manual
• ⚡ Unit test
• 💚 E2E

1. Create Deployment -

apiVersion: apps/v1
kind: Deployment
metadata:
  name: time-logger
  namespace: test
  labels:
    app: time-logger-dep
spec:
  replicas: 1
  selector:
    matchLabels:
      app: time-logger
  template:
    metadata:
      labels:
        app: time-logger
    spec:
      containers:
      - name: test-container
        image: containerlabs/aws-sdk
        command: ["sh", "-c"]
        args: ["while true; do for x in $(seq 1200); do date >> /var/log/time.log; sleep 1; done; truncate /var/log/time.log --size 0; done"]
EOF

2. Create Blueprint -

cat <<EOF | kubectl create -f -
apiVersion: cr.kanister.io/v1alpha1
kind: Blueprint
metadata:
  name: time-log-bp
  namespace: kanister
actions:
  backup:
    phases:
    - func: KubeTask
      name: backupToS3
      args:
        namespace: "{{ .Deployment.Namespace }}"
        image: busybox
        podOverride:
          containers:
          - name: container
            imagePullPolicy: IfNotPresent
        command: ["sh", "-c", "counter=1; while [ $counter -lt 20 ]; do echo $counter; ((counter++)); sleep 1; done"]
EOF

3. Create Actionset -

cat <<EOF | kubectl create -f -
apiVersion: cr.kanister.io/v1alpha1
kind: ActionSet
metadata:
  generateName: s3backup-
  namespace: kanister
  labels:
    kanister.io/JobID: hello-world

spec:
  actions:
  - name: backup
    blueprint: time-log-bp
    object:
      kind: Deployment
      name: time-logger
      namespace: test
EOF

4. Watch ephemeral pod and check its label -

NAME                     READY   STATUS    RESTARTS   AGE   LABELS
kanister-job-j298f   1/1     Running   0          0s    createdBy=kanister,kanister.io/JobID=hello-world

[viveksinghggits]

@mabhi are we just going to handle KubeTask in this PR?

[mabhi]

@mabhi are we just going to handle KubeTask in this PR?

Yes. Accounted additional pods as suggested in a separate PR.

[PrasadG193]

LGTM, minor suggestions about formatting and optimization in the tests.

[PrasadG193]

Added suggestions about additional validation we can add in the test.