Retry jobs prone to infrastructure failures

Igorrequested to merge
ig-retry-infra-failures into main

What does this MR do?

Replace the default runner_system_failure-only retry with an unconditional retry:2 mixin applied to job templates that are affected by transient infra failures (package, test, k8s integration, and all jobs on kubernetes runners).

The default no longer retries anything, so jobs that don't opt in aren't affected.

Why was this MR needed?

See #39338 (closed)

What's the best way to test this MR?

What are the relevant issue numbers?

Edited by Igor

Merge request reports