Frontend and pollers in different "locations"

I would like to know if anyone has tried to install librenms on aws with pollers in the datacenters and offices? would the performance be a problem? thanks!

Yes, but you have to manage latency.

Between the device and the poller you only have snmp latency.
You also have latency between the poller and the database and rrdcached server.
The database is the most latency sensitive, since it requires several round trips to update one piece of data.

distributed polling is not designed for remote polling, but it can do that in some cases.

It is probably best to put all LibreNMS parts in aws and have it send snmp requests over a tunnel to your datacenters.

that is the the one I can´t do…security people wont allow public cloud to reach our ouf of band network. (politics)
understood the compromises. thanks!

I don’t understand how you expected it to work? Probably best to turn up a server that is only on the out of band network.

my idea was to do snmp/cleartext polling on the datacenter and “sync” in someway that data with a master on the cloud. like zabbix does in their proxy deployment? like https://www.zabbix.com/documentation/3.4/manual/concepts/proxy

sorry to pick this up so late, but following on documentation this is exactly what i would like to do Scaling LibreNMS - LibreNMS Docs pollers on the DC´s/offices and the rest on aws

how to install dedicated packages on each server? one just with poller one with WebUI one with API?
thanks