For the monitoring design we will have separate pollers for separate regions.
For HA, the plan is that the devices are pointing to one IP for polling and routing will happen based on the proximity/shortest path to the region.
However, if one of the pollers die, the idea is that list of pollers is sent to the next nearest poller, so that the devices can continue to be polled.
How can we get a list of devices from librenms for that region, and how can it be “pasted” to another poller for continual access.
I could get a list of devices via API call. Not sure how to “paste” that on a different poller.
I’ll also have to store that data outside of the server, so if I lose the server/poller, I can get the list for adding it to another poller.
I will have some sort of health check in place wherein the pollers are pinging each other to make sure they are alive.
Does anyone have any insights here.
Will I have to do DB replication? Does the list of devices live in the DB?
And if the DBs are replicated across pollers, will the second poller automatically start polling (due to same IP on both pollers) when first poller goes down?