Using Thread Data to Monitor Slow Execution of Services
Slow services are identified and debugged by checking the thread-level data captured over a series of snapshots. This section describes the common thread issues. It also explains how the thread operations are related to services. These services are executed either by users or by your ThingWorx solution using schedulers, timers, or other events.
For thread-level analysis, you should capture several snapshots. The snapshots show how a thread activity changes over time. In case of slow performance, generally, five to ten thread captures for a period of 5 minutes helps to identify the performance issues.
The Support Tool utility captures thread snapshots every 30 seconds. To enable this functionality, use the CreateWatchdogFile service or manually create a file called runstacktrace in the Support.Tools repository on the disk. After taking the thread snapshots, delete the watchdog file by running the DeleteWatchdogFile service or remove the runstacktrace file manually. It you keep the functionality running, several GB of thread data accumulates every few minutes causing disk issues.