Stay up-to-date on Kubernetes development in 15 minutes a week.
It’s Release Week!
Nominations are open for the Kubernetes Steering Committee election. Also, check the elections app to see if you are eligible to vote, and file an exception if you aren’t.
Next Deadline: 1.28 Released, August 15
We are still on schedule to release 1.28 this week. Tuesday, even! Pending last-minute delays.
Patch releases have been delayed until August 23rd due to conflicts with the 1.28 release schedule.
KEP 3850 - Backoff Limits Per Index For Indexed Jobs
This KEP extends the Jobs API to support indexes for jobs, where the backoff limit is per index. The job can continue executing even if some of its indexes are failing. Currently, the indexes of an indexed job share a common backoff limit. When the job reaches this common backoff limit, the job is marked as failed by the job controller even if there are indexes which are yet to run to completion. The current implementation is not efficient if the indexes are truly independent of each other. If one index fails, the entire job would be marked as failed.
The proposed solution involves adding a new policy where the backoff limit controls the number of retries per index of the job. All indexes of the job execute until they succeed or fail. A new API field to control the number of failed indexes is also proposed. This KEP was authored by Michał Woźniak and jensentanlo and is in alpha in v1.28
-v
flags, even if it has a logging configLast Week In Kubernetes Development (LWKD) is a product of multiple contributors participating in Kubernetes SIG Contributor Experience. All original content is licensed Creative Commons Share-Alike, although linked content and images may be differently licensed. LWKD does collect some information on readers, see our privacy notice for details.
You may contribute to LWKD by submitting pull requests or issues on the LWKD github repo.