As an OpenShift developer, I want our e2e suite to run as fast as possible. To that end, our current e2e test suite takes approximately 2 hours to run and we've kept raising the timeouts. At this point, I do not think we should raise the timeout any further. Instead, we should focus on writing our tests in such a way that they can cover more ground at once. This could be through consolidating and refactoring tests, ensuring that a test only lands on a single node, to making our test suite fully concurrent.

Below text was provided by Matthew Kenigsberg:

Inspired by conversations with zzlotnik@redhat.com and walters@redhat.com and https://github.com/openshift/machine-config-operator/pull/2886#issuecomment-1021477022

Parallelizing e2e tests would both speed them up and ensure we don't leak state between tests

A possible way of doing this could be spinning up extra workers with the Machine API.