I have noticed that some network failures (alarm ping) monitored by net_connect take too long to generate an alarm. And this is causing customers suspect a bad work tool.
I've read some posts from other colleagues dessert first like symptoms and want to hear from anyone who can share your settings for delving more about this probe.
I currently monitoring approximately 150 data link internet on a single hub.
My settings are:
Profiles: ~ 150
Check interval: 5 min
Timeout: 60 sec
Failed interval: 1
Version probe: 3.03
If I want to receive alarms with a shorter time, which you would use?
The recommendation is to use 5 min (check) and 10 sec (timeout), to avoid processing queues.
In the probe of the documentation says that the profile takes too long to run the test, I devio reduce the amount of threads. But....
But I did a test in the laboratory and be too concerned.
In a unique setting, with only a single robot, using the settings and work normally.
But when we reduce the time (check) and the timeout, he immediately presents profile to complement all failure.
I can go back to the previous settings, but the profile execution failure remains. Only when I disconnect the server's network card the fault disappears.
I would like to hear from you using the probe of net_connect, which configuration you use that works efficiently.