Skip to content

Allow re-running failed parallel jobs before all parallel jobs finish #37900

@vasco-cruz

Description

@vasco-cruz

What problem would this solve?

When a workflow contains multiple jobs running in parallel, and one of those jobs fails, it is currently necessary to wait for all other parallel jobs to finish before being able to re-run the failed job.
This creates unnecessary delays, especially in workflows with long-running parallel jobs.
It would be very useful to allow users to immediately re-run a failed job independently, without waiting for the remaining parallel jobs to complete.

Screenshots

  • Screenshot 1: A failed parallel job where the Re-run buttons are not yet available because other parallel jobs are still running.
Image
  • Screenshot 2: The same workflow 7 minutes later, after all parallel jobs have finished, where the Re-run buttons finally appear.
Image

This demonstrates that re-running a failed job is currently blocked until every parallel job completes.

What do you propose?

If a job fails in a parallel execution stage:

  • The failed job should become available for re-run immediately.
  • Other parallel jobs should continue running normally.
  • The user should not need to wait for all parallel jobs to finish before retrying the failed job.

Metadata

Metadata

Assignees

No one assigned

    Labels

    type/proposalThe new feature has not been accepted yet but needs to be discussed first.

    Type

    No type
    No fields configured for issues without a type.

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions