Logs are records of events generated by your application during execution. They capture detailed information about requests, responses, errors, and system behavior. In Trusys, logs serve as the foundation for monitoring, debugging, and evaluating application performance and security.Logs can be explored at different levels:
Traces – Individual log entries representing a single request or operation.
Sessions – Groups of related traces, often used for analyzing multi-step interactions like conversations.
The Traces section provides a detailed, granular view of individual log entries from your monitored applications. You can view all traces for projects where monitoring is enabled. The list can be filtered by time range or application and is sorted by the most recent logs first. The list also indicates if a log has been evaluated, and if so, provides a summary of the evaluation and the span count.Click on any trace to view more details, including:
Span Hierarchy – A hierarchical view of the operations within a single request, useful for debugging and performance analysis.
System and Health Metadata – Additional information about the system state and health at the time the log was generated.
Evaluation Details – Comprehensive results of the configured functional and security evaluations performed on that specific trace. If an evaluation has not yet been performed for a trace, you will see an option to evaluate it against your application’s monitor settings.
Run Evaluation – If the trace is not evaluated, click on “Run evaluation” to evaluate the trace and see the trace evaluation result.
Sessions are primarily used for conversational logs where a session represents a group of related traces. This is particularly useful for applications that involve multi-turn conversations or extended user interactions.Click on any session to view key aspects of sessions:
Trace Grouping – A session contains multiple traces that are logically connected, typically representing a complete conversation or user journey.
Conversational Metrics – For conversational applications, metrics and evaluations are performed at the session level rather than individual trace level, providing insights into the overall conversation quality and performance.
Session-Level Analysis – View comprehensive analytics across all traces within a session to understand user interaction patterns and conversation flow.
