How much overhead does the agent use?
It is hard to quantify this in the general sense. We don’t have any published numbers on agent overhead since it can vary widely based on things like the platform’s hardware and operating system as well as the number of metrics the agent is collecting and how often it collects them. However, you can use the agent itself to give you a pretty good number for your specific environment since the agent itself emits metrics that can be used to get you its overhead information. The 2 metrics of importance are: • CPU Total Time per Minute: This allows you to see the actual CPU time in seconds the agent has taken. Of course this includes other agent activities like response time collection, event/log tracking, etc, but it gives a good overall picture. • Total Time Spent Fetching Metrics per Minute: This is the wall clock time taken to collect metrics every minute. So for example, consider if you had a platform whose agent is monitoring System, JBoss, Tomcat, Apache, Postgres, etc. The agent’s m
