Measured Latency incorrect

Products

• CA eHealth

Issue/Introduction

Introduction:

It may be the case where you notice Latency in a report is much lower than expected compared to when you are manually pinging a device from the command line.

Background: 

When the poller starts the poll cycle, it pings the ip address 4 times, drops the first ICMP reply, averages the following next three ICMP replies, and stores this information in the database under Latency.

Environment: 
Windows
Redhat Linux
Solaris 10

Instructions:

By default the latency measurement calculations are set to Min(Minimum) by default.

You can adjust the way Latency is calculated by adding the  following environment variable on the Poller system to change it to average.

NH_POLL_PING_MEAS_TYPE

Specifies the algorithm used to compute latency values based on multiple pings: minimum, average, or maximum.
The algorithms associated with this environment variable are applicable only to ping measurements, not alternate latency measurements.

min (Default)
Records the minimum response time.

avg
Records the average response time.

max
Records the maximum response time.

To enforce this change, you must stop eHealth so that when the poller initializes, it will read in this variable adjustment.

nhServer stop

nhServer start

Additional Information:

How to set eHealth environment variables in a Windows environment
http://www.ca.com/us/support/ca-support-online/product-content/knowledgebase-articles/tec513393.aspx 

How to set eHealth environment variables in a Unix environment:
http://www.ca.com/us/support/ca-support-online/product-content/knowledgebase-articles/tec513391.aspx 

Environment

Release:
Component: EHPOLL