Home
Products
Network Performance Monitor (NPM)
Node Alert interval
mjhodge
What is the algorithm NPM usies to determine in a node in "Down", and what is the maximum time delay before such a determination?
I am using a standard polling interval of 120 seconds, therefore I would guess that in a worse case scenario the node fails at poll+1, the delay would be 119 seconds. However i'm not sure if multiple polls need to fail, or if there is some other algorithm in place to determine if the node is "really" down or if it is just a latency issue.
Find more posts tagged with
Accepted answers
All comments
Network_Guru
I did a search for
Fast Polling Timer
& found this detailed thread:
www4.solarwinds.net/.../topic.asp
You might also want to put a sniffer on your network to verify exactly how Orion polls unresponsive nodes, as the algorithm may have changed since this thread was posted.
-=Cheers=-
NG
mjhodge
Cheers, from this it seems the worst case scenario is that a node could be down for 239 seconds before a "Node Down" alert is generated. i.e.
Node failes at Poll+1 seconds, next poll 119 seconds latter.
Poll failes, Node goes into "Warning" state and fast poll beggins.
120 seconds fast polling fails, Node goes into "Down" state.
Dan.Wendeln
mjhodge,
You are correct in your logic. Orion was designed that at the first missed ping the node will go into the warning state and fast polling would begin. At that point the system is polling the device every 10 seconds for the duration of the fast poll interval. If the device responds to one of these polls then the device is marked back as up and leaves the warning state. Please note that the fast poll interval is configurable. The default is 120 seconds but can be adjusted via the Advanced Settings menu in the system Manager.
Dan Wendeln
SolarWinds.Net
Customer Support Services
Quick Links
All Categories
Recent Posts
Activity
Unanswered
Groups
Help
Best Of