How the net_connect probe alert settings work

book

Article ID: 34655

calendar_today

Updated On:

Products

DX Infrastructure Management NIMSOFT PROBES

Issue/Introduction



There are number of variables in the net_connect probe.This Article will give you a better understanding of the alert settings.

Environment

Release:
Component: UIMNCN

Resolution

  • Send burst of "a" messages
  • Timeout after "b" seconds
  • Interval "c"
  • Retries in interval "d"
  • Failed intervals "e"

The probe initiates a test on every interval "c". It sends a burst of "a" pings. Then, it waits for a timeout after "b" seconds. If it receives an echo, the test is complete.

If there is no response within the timeout after "b" seconds, the probe sends another burst of "a" pings. As long as there is no response, it does this again until it has retried "d" times.

If it has re-tried "d" times in the interval, then it checks to see how many times in a row this entire cycle has failed.

If it has failed "e" intervals, then it generates an alarm. Otherwise, it waits for interval "c" and repeats the test.


EXAMPLE:
your original configuration would have been.

a = 3
b = 45 Sec
c = 5 min
d = 5
e = 3

(Timeout * Retries) * Failed intervals < interval
(45 * 5) * 3 = 675 seconds or 11.25 Minutes

Interval = 5 minutes
the profile setup could fail as the time to finish is GREATER THAN the polling interval.


Your new values:
a = 3
b = 20 Sec
c = 5 min
d = 5
e = 3

(Timeout * Retries) * Failed intervals < interval
(20 * 5) * 3 = 300 seconds or 5 Minutes

Interval = 5 minutes
the profile setup could fail as the time to finish is EQUAL TO the polling interval.
it should be less.

Working sugegstion:
a = 3
b = 15 Sec
c = 5 min
d = 5
e = 3

(Timeout * Retries) * Failed intervals < interval
(15 * 5) * 3 = 225 seconds or 3.75 Minutes

Interval = 5 minutes
the profile setup should work fine as the profile time is LESS THAN the polling interval