During beaker maintenance we need to suspend qe-pipelines jobs to avoid errors, we should do similar to what CKI jobs do.

We should also herder-retry connection errors with beaker in qe-pipelines (cki pipelines shouldn't retry them, because their tests are broader and long running), e.g.:

• https://gitlab.com/redhat/red-hat-ci-tools/kernel/qe-pipelines/kernel-ci-pipelines/-/jobs/9657032348

AC:

• [ ] herder: retry beaker connection error in KQE-pipelines
• [ ] List KQE "gitlab runners" in https://cki-project.org/docs/cki/operations/shutdown/
• [ ] Mention "`adhoc` runners in Beaker" in https://cki-project.org/docs/cki/operations/shutdown/

Jira: CKI-6878(https://issues.redhat.com/browse/CKI-6878)