When you come across a situation where an ITM Agent is using an abnormal amount of CPU, Memory, Process threads, or some other type of system resource, there are few important things you may want to consider as you try to identify the root cause.
All agent architectures are comprised of two primary worker threads, one responsible for executing instructions from the Agent Framework layer also called TEMA, the second executes the agent data collection cycle. This applies to both OS and application type agents. The Agent Framework is responsible for generating regular heartbeats to the TEMS and also receives instructions from the TEMS for such things as a real-time data queries to populate the workspace view, or to evaluate a situation.
Any agent data query will run the worker thread to collect data. A query will involve a specific attribute group that will have many discrete data collection activities. The agent uses specific platform or application API calls to perform the data collection. This is where most of the resource consumption issues have been found.
There are three specific actions that run data collection thread:
1. Workspace update
2. Situation evaluation
3. History data collection
The key to identifying the root cause of the problem is by using the process of elimination. Start by suspecting the most recent activity. Perhaps a new situation that was introduced, or a new attribute group added to history data collection. If there are no obvious suspects, then you will need to investigate each one of them separately. Stop all situations, all history data collection, and workspace updates. Restart the agent to ensure a starting baseline. Then enable one activity at a time and observe. If the problem re-emerges, work to further isolate the problem to a particular workspace, a situation, or history data collection attribute group causing the problem.
Subscribe and follow us for all the latest information directly on your social feeds: