When using the ignore on-fail strategy like the example below, if a step does fail and is ignored, it's not possible to re-run that step:

on-fail:
  strategy: ignore

This is documented here, and is expected behavior:
https://www.atlassian.com/blog/bitbucket/introducing-step-failure-strategies-in-bitbucket-pipelines

However, there's no easy way to re-run this failed step. This feature request is to add the ability to re-run the step, even if it was ignored.