Sergeant

- The supervisor now waits for all its children that became zombies. This behavior happens when the process killer kills its watched worker with SIGKILL which replaces its parent to the supervisor. The supervisor didn't know it before, so no one would call to wait on its pid. I added a zombie cleaning method to run on each supervising iteration to call `wait` for all the zombies.

- process killer now kills and waits for child processes that were spawned by the process to kill

<h1>Change list</h1>

- <b>Timeouts mechanism overhaul</b>:
- Removed `soft_timeout`/`hard_timeout` configurations in favor of just `timeout`
- Removed `critical_timeout` in favor of a defaulted `grace_period`. The rationale was that nobody understood what is the meaning of "critical" timeout. The idea of the critical timeout was to allow the last resort option of killing the worker in cases where SIGTERM did not succeed. I find `grace_period` to be more convenient and intuitive.

<h1>Breaking Changes:</h1>

- Sergeant Timeouts config does not provide `soft_timeout`, `hard_timeout` and `critical_timeout` anymore. They were replaced with `timeout` and `grace_period`

- fixed a bug when the process killer was killing the worker while it was still running

- fixed a bug when the process killer hanged waiting for the watched process to interact when it was abruptly dead
- increase verbosity on supervised worker unexpected startup failure

- `LogstashHandler` did not define any timeout which caused connectivity problems to raise timeout exceptions after a lot of time. This behavior made bad Logstash connections slowing down workers tremendously.