I've come to the conclusion that these timeouts are doing far more harm than good. Far too many times I've been unable to retrieve diagnostic information from a customer's Nexus instance because of these automatic timeouts.
Note: I'm filing this here so we can discuss on Thursday, not putting into a release yet.
If we decide to do this they can be disabled by setting their values to "0".
Also marking for release note in case we proceed with this.