eHealth Live Exception - Variable Time Over Threshold Alarm showing up later than expected
Article ID: 107017
We are using eHealth 22.214.171.124 on a Linux machine. The element is a GigabitEthernet0/0 interface and is part of a group that is attached to an alarm profile with the following rule: - Element Type = LAN Interface - Rule Type = Variable - Severity = Critical - Rule Message = Test Lan Interface Threshold - Time = 1 Minutes out of 5 Minutes - Variable - Bits /sec - Condition Type = Time Over Threshold - Above Threshold of 100.0 Bits The Problem Start is showing as the time we set up the profile and rules at 5/2 2:06:08PM But the alarm is occuring at 5/2 2:21:07PM Why is there a delay of when we are seeing the alarm?
You have a threshold set for 1 minute out of 5, breach if over 100bits. And you have a polling period of 15 minutes.. Let’s say the poll starts 12 We hit the threshold at 12:01 But eHealth won’t get the data for this poll period until 12:15. Now add a few minutes to process the poll data, compare the threshold, trigger an alarm, and then hand the alarm to the notifier process which tells OneClick. So for a breach that occurred at 12:01, you won’t actually see the alarm until ~12:18 in the best case
Reducing the poll period to 5-minutes would reduce the possible lag time but would increase the load on eHealth.
adjust the poll period and assess the impact to the system, or apply the shorter poll to only selected elements.