The invention relates to an application monitoring tool. According to an embodiment of the present invention, the application monitoring tool has an integrated event/tracing API that traces distributed and non-distributed system flows. For example, a tracing approach may be integrated with a Metrics functionality in that tracing may be throttled or turned on/off based on defined intervals, threshold breaches, and/or anomaly detection of related metrics. If an application's web service calls are beginning to take longer than “normal,” an event/tracing collection feature may be activated to trace calls to enable deeper analysis of the problem.