shared runner: long queue and job not in parallel anymore?

Recently I started facing longer queue duration for the CI jobs (up to 4 hours) and it seems jobs don't run in parallel anymore, or not more than 2 or maybe 3 at the same time. This happens every now and then.

Was there some change with the shared runners? Is this due to congestion or a lot of jobs queued? Or is there some new limitation in runs (per user or repository)?

BTW is there a place where we can check the load of the shared runner?

(issue was originally posted here salsa-ci-team/pipeline#337 (closed))