Sometime in IBM Tivoli Monitoring (ITM), agents that are reported to be online are not working accordingly. Hence, situations are not running and you will not get alerts for the thresholds that you are monitoring. You fail to meet SLAs and this can be costly. It is almost impossible to find those agents by using an ITM provided mechanism. Such agents occasionally time out when real-time data is requested and the status will be set to offline, but not always. They are referred to as “not-responsive” agents in this resource.
Mostly you are alerted by your application administrators or customers that they did not get any event for their business application, even though an ITM situation was set up to monitor a specific condition. Usually, you check the agent’s health by using a Tivoli Enterprise Portal real-time data request, for example, a disk usage report. If the agent goes offline after such a request, you start it again. If it remains online, but no data is returned, your next action is to check the logs and restart the agent for quick problem resolution. After the agent restart, you will probably execute several tests to verify the agent’s behavior. All of the tests will be successful mostly and the issue appears to be solved. However, you are not certain because you do not know when and where the problem might reoccur.
This solution download helps you to monitor not-responsive agents and avoid manual work to find those agents beforehand. It can be used in two different ways, as a Standalone script or an Agent Builder Custom Agent.
How it works
This solution can be used in two ways.
- Script only mode: A script is executed manually. Results are written to STDOUT and files. Details can be found here in this wiki.
- IBM Agent Builder custom agent
The main part of this tool is a simple Perl script that does the following:
- Using SOAP to connect to the IBM HUB Server
- Creating an agent status list before real time data request
- Retrieve real-time data from the agents
- Creating an agent status list again
- Check for which agents no data have been returned and reporting those agents to STDOUT and into a file.
Details can be found here in this wiki.
Supported Operating System
The procedure was tested on Windows and Linux Redhat + Suse but should run on other UNIX and Linux Operating systems.
PERL needs to be installed on the computer where this tool is going to be used.
The following PERL packages are used and need to be available:
- >use LWP::UserAgent
- >use XML::Simple;
- >use Getopt::Long;
- >use Data::Dumper;
- >use File::Basename;
- >use Time::Local;
- >use Sys::Hostname;
Released: 03 April 2019
Cost: No Charge
Support: Not Supported
This package is subject to the terms and conditions displayed upon download.
This download has been prepared and reviewed by IBM. It is offered “AS IS” to existing customers. Support is not provided for this download.