review-apps - Ensure that review-deploy retries don't block one another

Context

In !114444 (merged), we started retrying review-app CI jobs a few times.

It helps for some problems (see !115490 (comment 1323809913)), but it creates other issues where the retry jobs don't go too far, because we are cleaning up the review-app environment in the meantime (e.g. https://gitlab.com/gitlab-org/gitlab/-/jobs/4054042914):

WARNING: Kubernetes configuration file is group-readable. This is insecure. Location: /builds/gitlab-org/gitlab.tmp/KUBECONFIG
WARNING: Kubernetes configuration file is world-readable. This is insecure. Location: /builds/gitlab-org/gitlab.tmp/KUBECONFIG
Error: UPGRADE FAILED: another operation (install/upgrade/rollback) is in progress
[21:04:29] Retrying 2...
WARNING: Kubernetes configuration file is group-readable. This is insecure. Location: /builds/gitlab-org/gitlab.tmp/KUBECONFIG
WARNING: Kubernetes configuration file is world-readable. This is insecure. Location: /builds/gitlab-org/gitlab.tmp/KUBECONFIG
Error: UPGRADE FAILED: another operation (install/upgrade/rollback) is in progress
[21:04:33] Retrying 1...
WARNING: Kubernetes configuration file is group-readable. This is insecure. Location: /builds/gitlab-org/gitlab.tmp/KUBECONFIG
WARNING: Kubernetes configuration file is world-readable. This is insecure. Location: /builds/gitlab-org/gitlab.tmp/KUBECONFIG
Error: UPGRADE FAILED: another operation (install/upgrade/rollback) is in progress

Goal

Ensure that the retries don't overlap on one another, or remove the retries altogether.