Change to sensible timeouts

[jvalkeal]

Trouble with central is that it is flaky and current timeouts are not good. We need to find a balance where defaults simply work 99% of a time while bailing out from a workflow if things just hang too long.

In recent months when central has had a bad day, operations will error, close may take up to 2 hours, release 1 hour among other issues where api is just broken.