🐛 Describe the bug

Submitting datainputs via:

kubectl apply -f foo.yaml

results in modified job names that can exceed an unexpected 63 character limit. When this happens, the job will sit inactive without complaint.

One has to know to look at:

kubectl get datainput <job name> -n <namespace> -o yaml

and look for this magic:

      \\\"<job name>\\\"
      is invalid: [metadata.labels: Invalid value: \\\"m2020-datainput-ids-mars2020-edlcam-ops-calibrated-browse-ddc-datainput\\\":
      must be no more than 63 characters

Some of the name was extended after the data input was applied, so the expanded length was also unexpected:

echo -n m2020-datainput-ids-mars2020-edlcam-ops-calibrated-browse-ddc | wc -c
61
echo -n m2020-datainput-ids-mars2020-edlcam-ops-calibrated-browse-ddc-datainput | wc -c
71

🕵️ Expected behavior

It would be great if this error could fail loudly when kubectrl apply is used and logged such that the user could take action

🩺 Test Data / Additional context

apiVersion: kdp.nasa-pds.github.io/v1
kind: DataInput
metadata:
  namespace: m2020-kdp
  name: m2020-datainput-ids-mars2020-edlcam-ops-calibrated-browse-ddc
spec:
  pipeline:
    id: product-validation-pipeline
  data:
    bundle_bucket: s3://<bundle-name>
    bundle_subpath: <prefix_to_sols>
    out_dir: s3://<another-bundle-name>
  image: <über-cool-image>.amazonaws.com/kdp-datainput-prefix-listing:latest

🖥 System Info

• OS: AWS AMI