the problem here is exacerbated by the fact that each indicator chart has an individual refresh internal, which defaults to 60 seconds. this wouldn't be so bad if the refresh countdown began AFTER each chart finished rendering, but that's not how it works. instead, the 60 second interval is mapped to the clock on the wall, so if it takes 55 seconds to generate the graph, it will refresh 5 seconds later. if it takes longer than 60 seconds to generate the graph, it will put constant load on the database without ever generating anything (because by the time one chart is generated, it's auto-refreshing-div element has already requested new data and discarded the previous request).
rev2406 - do not load the method that registers the refresh callback;
QA Verified, we're no longer refreshing the chart every minute or so, in the CP build.
This bug was previously known as http://jira.rhq-project.org/browse/RHQ-1265
Mass move to component = Monitoring