Performance Data Model and Content

Performance data is stored in a single, denormalized, repository table named TRON_PERFORMANCE, which you can query to collect performance data.

Each execution of project or job is identified by a global unique ID. Additional information is provided about the account that initiated the execution and the repository in which the project or job is located.

Note:

The starting time for execution and events are logged in Coordinated Universal Time (UTC), also referred to as Greenwich Mean Time (GMT).

For each executed project ID, version (modification date), name, and a global, unique execution ID is reported. A single project execution event stores the duration of a project in ms.

Project components are represented by ID, name, class, type, and version. The process event provides the number of steps and the amount of processed records.

For port events, the ID, name, class, type and the amount of input or output blocks and records are reported.

For each job, the ID, the version (modification date), and the name is reported. A single job execution event stores the duration of a job in ms. The components (projects) of a job are represented by their ID, class, and version.

Events

The performance log is based on events. For each event, the starting time and the duration (in ms) is stored. The description of an event is made up by a class, a name, and a text. Some events include a result (such as succeeded or failed). Also included is information about the engine that reported an event.

Reported Events
Class	Name	Description
control	execute job	Total execution time of a job (one record per job execution, job duration in ms, in attribute PRF_JOB_DURATION).
init	load job	Get job definition from repository.
control	execute project	Total execution time of a project (one record per project execution, project duration in ms, in attribute PRF_PRJ_DURATION).
init	load project	Get project definition from repository.
init	create	Create project and component instances.
init	configure	Configure project and component instances.
perform	prepare	Perform component preprocessing.
perform	process	Perform component step.
perform	finish	Perform component postprocessing.
perform	read	Get data to component input port.
perform	write	Push data from component output port.

Note:

Due to distributed multithreading, the total project execution time can be significantly shorter than the sum of the execution time of all participating components.