Polling Engine awareness

wrecktangle

Environment: NPM 12.2, NTA 2.2

I have multiple polling engines that have differing tasks/polling rates -- VIM, Routing, UCS, Wireless -- spread out amongst them.  However, all polling engines have Polling and Hardware Health.

The Question:

Is there any dependency between a network device and the specific associated tasks of a polling engine?  Specifically, do I need to add all polling engines to the switch/router ACLs for Solarwinds SNMP access?

Or, does the network device simply need access to any polling engine in order for NPM to properly communicate with the network device?

I would never do this, but for clarity sake, if only a single non-primary polling engine has SNMP access to all my network devices, will NPM/NTA still be able to capture all the required data from the network devices?

Thanks

- Jai

mesverrum

Each node is tied to one single poller, all tasks for that node will come from the same poller. 

At a minimum you need access for the one poller you use to monitor the node, but I find it is a best practice to add all possible pollers (except things like pollers on another network that couldn't reach the node to begin with) to the ACL because in the future you may need to rebalance the loads on your pollers, and if everything is locked down to just one poller you will not be able to do that without changing the configs on the devices as well.