In the case of my case in which im testing HQ. I do plan to get the enterprise version sooner or later. So for testing i installed the server and the agent on a VPS inside. Basically only this VPS runs inside the node.
Then I installed other 2 remote agent on 2 remote servers.
Basically, this is how i understand something. Please correct me if im wrong. At first to save some Memory i decided that to start the agent inside the VPS where the server also was is something useless, since the 2 other servers have their own agent. But then again i was realizing something, for network checks like HTTP, FTP etc, the documentation says it will use the agent where you create the New Platform server. So I though its not a good idea to monitor HTTP from the agent that is installed inside the same server you are going to monitor HTTP. The logic thing would be to do it from a remote server, so i started the agent. In this case i would have to use the agent installed inside the VPS to monitor network services for the 2 other remote servers. Correct? That leaves me that i basically dont need the other 2 agents installed. Unless of course i want some metics about cpu, swap etc.
Now if this is all correct, i was thinking if this main agent, like i will called since it will do the job to monitor network checks, if doing all the job for remote servers, why should it be inside the VPS? Would it not be better to install it on the same node where HQ runs but directly into the hardware instead of inside the VPS? This way i could also have data about the hardware node itself and doing network checks that are then send to the HQ inside the VPS. Im now confused on what would be actually better, or should i just leave it inside the VPS where the HQ server resides as well for better performance?