Fix noaccess retries

[pophilpo]

This PR will implement breaking out of the thread loop if the exit code of the port-forward call is 1 (#11). Before breaking out it will send a ChildEvent::Exit with a new RestartPolicy::NEVER

[pophilpo]

@sunsided At the moment this change will stop every process that exited with status code 1. I'm working on recoverable errors now but I think it should be another pull request

[sunsided]

It seems the newly added if statement here is either unnecessary or incorrect. Unnecessary because positive delay times are already captured above, incorrect because it leaves out the hypothetical negative delay times ("less that NONE"). What was the reason for adding it?

[sunsided]

At the moment this change will stop every process that exited with status code 1.

@pophilpo The restarting is integral to the functionality. If a pod is temporarily unreachable (e.g. due to a rolling update or scale-down event), we need the connection to be retried.

[sunsided]

@pophilpo Just added crashie to help testing that. The main branch here has a kind folder now that contains a kind cluster definition with a crashie deployment that will be restarted randomly. It has an HTTP port, so testing becomes a bit easier.