`supervisord-event-handler.py` is signaling that virtiofsd is stopping when it is actually still starting up

[Ramblurr]

There are two minor bugs in supervisord-event-handler.py

1. it is signaling that virtiofsd is stopping when it is actually still starting up:
2. its expected_count is off by one.

https://github.com/astro/microvm.nix/blob/f30ae9b9e740116d044935c9c0c68bd1c2ce8432/nixos-modules/microvm/virtiofsd/supervisord-event-handler.py#L39-L42

I added some extra debug to the python script, so you can see it:

Parsed event: PROCESS_STATE_STARTING
count=0/5 STOPPING
Parsed event: PROCESS_STATE_STARTING
count=0/5 STOPPING
Parsed event: PROCESS_STATE_STARTING
count=0/5 STOPPING
Parsed event: PROCESS_STATE_STARTING
count=0/5 STOPPING
Parsed event: PROCESS_STATE_STARTING
count=0/5 STOPPING
Parsed event: PROCESS_STATE_RUNNING
count=1/5 Process state running...
Parsed event: PROCESS_STATE_RUNNING
count=2/5 Process state running...
Parsed event: PROCESS_STATE_RUNNING
count=3/5 Process state running...
Parsed event: PROCESS_STATE_RUNNING
count=4/5 Process state running...
Parsed event: PROCESS_STATE_RUNNING
count=5/5 Process state running...
count=5/5 READY
Parsed event: PROCESS_STATE_RUNNING
count=6/5 Process state running...
count=6/5 READY

The off-by-one is because the handler's process itself is not taken into account, it is a simple fix, just add one to the expected count.

As for the stopping bug, I think the call to for stopping could be removed entirely. This python script isn't in a position to know if the service is stopping or not, all it can report on is whether the expected virtiofsd procs are ready.