Bump test timeout to handle long queues

We encounter some test jobs getting killed by timeouts due to longer Beaker queues. To avoid loss of testing results, increase the test timeout.

The scope of this change might be too wide and e.g. we might need to only bump the timeout for rhel8 or for s390x, but we couldn't find the necessary data to decide that quickly.

Even though we're trying to recover from those in the reporter right now, we really shouldn't as that complicates the code throughout the whole pipeline. Instead we should either increase the timeouts if such time is considered acceptable, or deal with queue length in Beaker.