Shutdown instances after maximum number of jobs

[DanielHeath]

Whenever there's enough load over a few hours to keep an instance around, it'll eventually start failing jobs because it's run out of some resource or another - most recently I've had ongoing issues with docker-compose not creating networks (even after upgrading to 5.8), but in the past I've run out of disk space too.

Being able to specify "Run X jobs / for X minutes, and then shut down" gives teams a quick work-around for resource leaks causing build failures.

[jamestelfer]

This would also be useful from a security point of view: ensuring that no instance lasts very long decreases the potential of a breached instance.