The current design provides a way to measure event cost, via "Enable Event Performance Logging" and Canister Log Level = Debug. However, this process requires a service restart and produces very large event process log files... for example 3GB within 10 minutes.

This request is to provide for suppressing debug-level messages when event performance logging is enabled... as those detail records are not needed in this situation and cause these files to be over-large.

Additionally, it would be helpful if high-level metrics would be provided by default, without this debug-level processing. For example, a periodic entry (hourly?) in the event processor and session closer logs with a total and percent 'minutes execution time'. The idea is to provide an inexpensive measure of total time evaluating an event, and perhaps a separate list of total time evaluating each hit/step attribute. The list might be limited to the top 100, etc., entries.

The goal is to make evaluation of event cost easier to perform, so that costly events can be re-designed or disabled.