Take a look at your poller completion. Depending on how many nodes you have you may be overtaxing the status poller. Take a look at Settings -> Details -> Polling Engines. If those are OK yau can reduce the warning time some. One problem is that you may start to get false positives for devices which become unreachable from PM for a short period of time. Ypu'll have to adjust the numbers and look for false positives or overloaded pollers.
Thank you for your advice.
- Default Node Poll Interval : 60 seconds
- Node Warning Level : 30 seconds
In this case, it seems Orion NPM v10.0 handles as follows (in our environment):
(Devices stop responding) --> ICMP Ping -- (60 seconds) --> ICMP Ping -- (60 seconds) --> Start Fast Polling (Node Status is Warning) -- (30 seconds = 10 seconds * 3 or 4 times) --> Node Status is Down.
*It takes 2 minutes for Orion NPM to change the node status from UP to Warning.
If it were not for the second ICMP Ping, it would be easier for us to configure the polling interval and the node warning level.