Backoff Limit Per Index For Indexed Jobs

[jensentanlo]

Enhancement Description

• One-line enhancement description (can be used as a release note): Add New Indexed Job backoff limit mode that is counted per index rather than per job

• Kubernetes Enhancement Proposal: https://github.com/kubernetes/enhancements/tree/master/keps/sig-apps/3850-backoff-limits-per-index-for-indexed-jobs

• Discussion Link: An Indexed Job mode that allows every index to execute kubernetes#109712

• Primary contact (assignee): mimowo

• Responsible SIGs: apps

• Enhancement target (which target equals to which milestone):

  • Alpha release target (x.y): 1.28
  • Beta release target (x.y): 1.29
  • Stable release target (x.y): 1.33

• Alpha

  • KEP (k/enhancements) update PR(s): Backoff limit per Job Index #3967
  • Code (k/k) update PR(s):
    • Support BackoffLimitPerIndex in Jobs kubernetes#118009
    • Extend the Job API for BackoffLimitPerIndex kubernetes#119294
  • Docs (k/website) update PR(s): Docs update for Job's backoff limit per index (alpha in 1.28) website#41921

• Beta

  • KEP (k/enhancements) update PR(s):
    • Update KEP-3850 "Backoff Limit Per Index" for Beta #4228
    • Update KEP3850 "BackoffLimitPerIndex for Indexed Jobs" #4321
  • Code (k/k) update PR(s):
    • Graduate BackoffLimitPerIndex to Beta kubernetes#121356
    • Introduce the job_finished_indexes_total metric kubernetes#121292
    • Indicate Job Backoff Limit Per Index reason consts are beta kubernetes#121471
    • Backoff limit per index e2e test kubernetes#121368
    • Add remaining e2e tests for Job BackoffLimitPerIndex based on KEP kubernetes#121633
    • Benchmark job with backoff limit per index kubernetes#121393
  • Docs (k/website) update(s): Docs update for Job Backoff Limit Per Index in Beta website#43388

• Stable

  • KEP (k/enhancements) update PR(s): KEP3850: graduate Backoff Limit Per Index for Job to stable #5154
  • Code (k/k) update PR(s):
    • Graduate Backoff Limit Per Index as stable kubernetes#130061
    • Add Job e2e for tracking failure count per index kubernetes#130390
  • Docs (k/website) update(s): Docs update as we promote JobBackoffLimitPerIndex to stable website#49811

Please keep this description up to date. This will help the Enhancement Team to track the evolution of the enhancement efficiently.

[jensentanlo]

/sig apps
/wg batch

[alculquicondor]

/assign @mimowo

[alculquicondor]

In addition to configuring the backoff per index, we should probably have FailIndex as one of the actions for pod failure policies.

[soltysh]

/milestone v1.28
/stage alpha
/label lead-opted-in