Analyzing PSM PurePaths to Monitor Slow Execution of Services
PSM records service execution metrics for some select operations over time. By default, PSM captures the data on http threads. Therefore, any bottlenecks in the Event Processing subsystem or other ThingWorx subsystems are not collected in this tool. However, slow user operations are visible, and you can use PurePaths to determine which services and user actions are long running and should be optimized.
The Dynatrace PurePath technology records all user transactions and interactions performed on the ThingWorx server. PurePath records thread-level details that show how long each method took to execute. Additionally, PurePaths also record the duration for executing internal methods and shows how many resources each internal method consumed. PurePath also breaks down the wait time based on where the wait occurs, for example, slow IO, CPU, JVM suspension, and so on.
PurePaths give a direct view in the slow user transactions over time. For example, the following default PurePath dashboard enables you to sort services on the total execution time:
In the example, you can see that several services are performing slowly. These services impact the users. The users have to wait for over two minutes for a mashup service to complete. PSM captures the end-user experience. It does not collect data for timers or schedulers that may impact backend processing.
Check the response time to identify services that need optimization. Dynatrace also helps you identify the longest API call in a service. In this example, the longest API call is QueryImplementingThingsWithData.
Was this helpful?