Docs: Enhance the distributed monitoring chapter with agents, satellites and masters

Docs: Enhance the distributed monitoring chapter with agents, satellites and masters

Next to renaming client->agent, I've also refined:

- the health check chapter follows the two scenarios above and provides more concrete configs
- log_duration=0 for agent endpoint objects
- advanced scenarios don't manage the agent zone/endpoint in zones.conf anymore, but zones.d/master
- Ensure that the connection direction is correct inside the scenarios. Avoid doubled/hanging connections.