The CA Standard Monitor offers an option to limit the impact of application keep-alive messages on monitoring statistics in reports. The technique involves limiting a selected application’s Server Response Time (SRT) or Data Transfer Time (DTT) to a maximum value so that unnecessary SRT or DTT observations are ignored. You can set the value to a number of seconds that falls just below the keep-alive frequency.
If you suspect that an application is sending keep-alives, look for an inverse relationship between observations and SRT, and for SRT averages in the second range instead of the millisecond range. Configure the CA Standard Monitor to limit the maximum SRT for an application.
If you determine that your application uses keep-alives that result in extremely high Data Transfer Times (DTT), you can also apply a similar limit to filter DTT.
Use SRT and DTT filtering on the CA Standard Monitor with the Console’s built-in NRTT filtering. Typically, application keep-alives skew SRT data when clients are idle. NRTT filtering is useful at times when all clients are idle, such as non-business hours. However, during the day you can easily have a client network that has 10 idle connections and 30 active connections. You could surpass the NRTT threshold but still have skewed data due to the 10 idle connections inside that combination.
If you are unsure of the keep-alive frequency of a selected application, we recommend using 10 seconds as a safe starting point. It is highly unlikely that a server would take more than 10 seconds to start responding to a user request. In most (but not all) cases, the keep-alive frequency will be greater than 10 seconds.
For applications that use random ports, such as Microsoft Exchange 2007, the easiest way to identify the port is to open Outlook, run a netstat command on your computer, and record the dynamic port to which Outlook connects.
The required setting is also available on the CA Multi-Port Monitor. However, the steps to take to change the default setting are different. For more information, see the CA Multi-Port Monitor product documentation.
Follow these steps:
The LimitServerResponseParms.ini text file accepts multiple entries, separated by line breaks. Limit the SRT of each application by supplying the port number and the maximum amount of SRT that is allowed. Set the maximum SRT value to a value that is slightly less than the keep-alive frequency. For example, to ignore Citrix keep-alives that are occurring at a frequency of 60 seconds, supply the following entry:
/port=1494 /max seconds=59
/min port=<lowerPort> /max port=<higherPort> /max seconds=59
If no port is specified, or is specified as being 0, then the specified limit is applied to all ports. A port range with 0 as its lower limit will have the same effect, regardless of what the upper port limit is specified as.
Entries that appear earlier in the file have priority over entries that appear later. Because of this, if there are various rules with overlapping port ranges, it is imperative to list the more specific ones first, otherwise they will be masked by the less specific ones. For example:
/port=23 /max seconds=15
/min port=100 /max port=200 /max seconds=50
/max seconds=120
This file limits SRT or DTT to 15 seconds for port 23, to 50 seconds for ports [100-200] and to 120 seconds for all other ports.
, then choose Synchronize Monitor Devices.
|
Copyright © 2014 CA Technologies.
All rights reserved.
|
|