Cases in Pega Predictive Diagnostic Cloud
Pega Predictive Diagnostic Cloud™ (PDC) gathers alert messages, exceptions, and other data that Pega Platform™ provides, and then generates events and organizes them into cases. To better understand how that data helps you track and monitor issues in your system, learn how PDC creates cases.
PDC cases help you analyze issues, as well as assign them and track their resolution. For more information, see Managing cases in PDC.
Pega Platform provides PDC with various types of information, such as alerts, exceptions, and performance statistics. Events are data points that PDC creates based on that information.
PDC supports the following types of events:
- Alerts – Messages that Pega Platform generates to report performance and security issues, for example, the PEGA0001 alert when an HTTP interaction lasts longer than the threshold setting. For more information, see Alerts.
- Exceptions – Errors that interrupt the operation of your application. Pega Platform includes exceptions in the PegaRULES log file, which PDC interprets to identify specific issues. For example, when an application cannot allocate an object in the Java Metaspace because the Metaspace has run out of memory, Pega Platform generates a java.lang.OutOfMemoryError exception, for which PDC creates the OPS0025 event.
- Insights – Observations that PDC makes based on the activities and changes in your system, which help you diagnose and resolve issues. These events are different from alerts and exceptions because they provide additional information that is the result of a synthetic analysis of your system. For example, when an agent is in the Exception state, PDC generates the OPS0015 event.
Events gather a wide range of data, such as session descriptions, stack frame lists, clipboard parameter page data, and the Performance Analysis (PAL) counters of a session at the time of an event. PDC receives this data and then stores the records in a database. For more information, see Data collected by Pega Predictive Diagnostic Cloud. For more information about each event, see List of events and notifications in PDC.
Every five minutes, a PDC agent evaluates all events in your application since the last run of the agent, and then groups together the events that resulted from the same issue. Such a grouping by issue is a correlation. PDC correlates events to create cases that represent specific issues. For example, to correlate a slow query alert with the appropriate case, PDC identifies the query, and the rule that defines the query. To correlate exceptions that originate from the same part of your application, PDC determines the class where each exception occurred, and then parses the exception messages.
If PDC does not find any cases from a specific correlation, it creates a new case with a type that identifies the problem, and then updates the case when additional events from the same correlation occur.