Ton of targets behind NAT - Suggestions

So, I have to monitor over 500 (and growing) targets windows machines all around the globe.
It’s like 10-20 targets per location. Zero of the firewalls there are under my administration, so can’t open ANY ports.

I already have remote access to all of them with RMM tools.
So I can script install anything I want.

Real problem/question

I plan to have a VM in a cloud provider for Prometheus server (most likely in docker, with traefik reverse proxy).

As for the targets, I’m thinking on having on each one:

  • The exporters I want (wmi_exporter, smartctl_exporter, and a mysql exporter on few of them).
  • A prometheus in agent mode, with remote write to the prometheus server in the VM

I consider having 1 prometheus agent per network, but targets are not 24/7 up, all are individually going up and down based on the usage of the store.

The question is, will a single prometheus server handle the incoming traffic of 500 prometheus agents?
(Assuming we have a time that all of those are online).

Any recommendation on what else I could do?