- Count all time the unhealthy events, if threshold is passed and service is in started then stop it.
- Better loggig facility. Using a file as stdio it's not the best.
- Send SIGKILL to whole processgroup when killing a service
- Stack all the things!
- Parameter "start-if-failed". Might be worth it to generalize start-if = [ServiceName: Status]?
- Create another binary for getting the status of the services:
- Send ServiceAdded event and handle runtime services addition
- Services config file validation
- Store timestamp when starting a new process (for knowing uptime)