SAM Polling Rate after update to NPM 11.5 / SAM 6.2.0

It's hard to get a sense for the number of component being monitored from poller details view. Can you post a screenshot of the SAM License Summary View [Settings -> SAM Settings -> SAM License Summary]

Allowed Number of Component Monitors: unlimited

Total Number of Component Monitors: 21490

  Licensed Component Monitors: 21490

  Unlicensed Component Monitors: 0

Available Component Monitors: unlimited

I got it to drop to 70% by deleting just about everything. Still not quite making sense to me. Doubling the cores should have dropped that value significantly just like it did on the other pollers.

How many if any additional pollers do you have? Each poller (including the primary) can support upwards of 8-10k components. 20,000+ components would require at least one Additional Poller.

We have 4. I did a lot of housekeeping and usage is almost nothing on most of the boxes, except for that same questionable box.

1. This is another remote site within 8ms latency of the poller in question.

2. This is the poller that is over representing its load.

3. This is our primary sever. We are in the middle of a failover right now so the low polling completion is fine.

4. Additional server for primary site.

It looks like the overwhelming majority of the servers you are monitoring with SAM are on the 2nd poller listed above, with the 70% polling rate. The other three pollers are hardly utilized from a SAM perspective. Instead of deleting/removing items from monitoring you can move some of those those server nodes from the 2nd poller to your other polling engines. This will help distribute the load more evenly across your four pollers.

The Short:

Try confirming if the polling intervals are whats expected by going to 'Edit Properties' on an individual node or interface. 

The Long:

Not sure if this was the same issue or not, but I ran into something similar today with NPM when changing my polling intervals from 60sec to 30sec using Orion Polling Settings.  I updated both 'Default Node Poll Interval' and 'Default Interface Poll Interval'.  It appeared to work fine as I saw my polling rates spike on all three pollers to include one that went over 100%.  I then switched back to 60sec polling intervals and saw no decreases in the polling rates.  After monitoring for a while to see if the rates would drop, they had all remained the same.  I then went into my 'Edit Properties' and found that the individual nodes were still showing 30sec, even though my Polling Settings were set to 60sec.  After updating the polling rates on all of my individual nodes and interfaces using 'Manage Nodes' I then saw polling rates come back down to their expected levels.

This is also my problem....

How can I check if how many components are already used in each poller server? Thank you.

FYI. Just created the report that will display the application name, number of components, and Pollers

hi aLTeReGo‌, so the maximum components per poller are 8-10K components only?

Correct. That is the recommended limit per-polling engine. SAM 6.2+ supports upwards of 150,000 components per-Orion instance. Stacked pollers can be used to achieve 20,000 components per-server.