Cherry-pick of PR #20305 to 7.8 branch. Original message:

What does this PR do?

This PR ignores ErrInputNotFinished error occur in autodiscover stop/start process. This is required in order to avoid having stoped configs that never come back since the start event fails due to this error on first attempt if the previous state is not cleaned yet.

Currently configs that fail due to ErrInputNotFinished are skipped at

Why is it important?

In order to resolve a permanent issue with updated Pods, which makes Filebeat stop collecting logs after a Pod is updated.

How to test this PR locally

1. Deploy Filebeat on k8s using the following config for autodiscover (set a valid output too so as to ship logs to ES):

filebeat.autodiscover:
  providers:
    - type: kubernetes
      node: ${NODE_NAME}
      templates:
        - condition:
            equals:
              kubernetes.pod.name: "mytarget3"
          config:
            - type: container
              paths:
                - /var/log/containers/*${data.kubernetes.container.id}.log

2. While Filebeat is up and running deploy a target pod to be autodiscovered and make Filebeat collects its logs:

---
apiVersion: v1
kind: Pod
metadata:
  name: mytarget3
  labels:
    app: test
spec:
  containers:
    - name: test
      image: ubuntu:latest
      command:
        - bash
        - -c
        - |
          #!/bin/bash
          echo "$(date): started the process"

          while :
          do
                 echo "$(date): sleeping 5 seconds"
                 sleep 5
          done

3. Update the target Pod's manifest by adding an extra label like team: qa
4. apply the Pod's update with kubectl apply -f <manifest_filename>.yml
5. Make sure that after a while, Filebeat continues collecting logs after the update of the Pod.