Automatically retry integration_k8s jobs

What does this MR do?

Adds automatic retries to the integration_k8s job.

Why was this MR needed?

Automatically retry integration_k8s jobs

The job have known flaky tests that we haven't yet fixed. It's already annoying when it fails the pipeline (especially the merge train ones!) because of that. And given our current experiment on Continuous Delivery of GitLab Runner to our maintained fleet (#28713) it becomes even more annoying.

We already use retry: 2 in few of our jobs that are known from failing randomly on unknown reasons. This commit follows the same approach.

Merging this doesn't of course mean that we shold not identify and fix what's causing integration_k8s random failures. It just aims to make the job less annoying for the maintainers until this is done.

What's the best way to test this MR?

What are the relevant issue numbers?