[CI] Pipeline failures for `review_gke122`
Summary
Likely due to auto-upgrade of gke122
cluster to 1.25
we're seeing quite a few failures in pipelines.
We need to either:
- re-create
1.22
cluster and "pin" version hard (it's troublesome due to current GKE limitations , see Outline our CI cluster maintenance/provisioning... (gitlab-org/distribution&25) for more detail) - at the very least add
allow_failure: true
to existing*gke122
jobs to unblock pipelines. - remove
*gke122
tasks as with auto-upgrades we can no longer be guaranteed that version we deploy cluster at is static. Whereas*gke
can be the "floater" versioned cluster.
Edited by Dmytro Makovey