Aggregated token usage and function calling metrics by operation type.
Maps each event type within the phase to its complete aggregate metrics, including detailed token consumption breakdown with cache statistics and comprehensive function calling metrics data. This comprehensive aggregation enables deep analysis of resource utilization patterns and operation quality across the entire phase.
The partial record structure reflects that not all possible event types may occur during phase execution. Only operations that were actually performed will have entries in this mapping.
The aggregate data supports cost analysis (via token usage), reliability assessment (via function calling metrics), and optimization opportunities (via cache hit rates and failure patterns).
Timestamp when the event was created.
ISO 8601 formatted date-time string indicating when this event was emitted by the system. This timestamp is crucial for event ordering, performance analysis, and debugging the agent workflow execution timeline.
Format: "YYYY-MM-DDTHH:mm:ss.sssZ" (e.g., "2024-01-15T14:30:45.123Z")
Total elapsed time for the phase execution in milliseconds.
Measures the wall-clock duration from phase start to completion, encompassing all agent operations, self-healing spiral loops, compiler validations, and any retry attempts. This metric provides visibility into phase-level performance and enables identification of bottlenecks in the waterfall pipeline.
The elapsed time includes both active LLM processing and any overhead from compilation, validation, and orchestration logic. For detailed breakdown of time spent in specific operations, consult the individual operation events within the phase.
A unique identifier for the event.
Final state machine step counter value for the phase.
Records the terminal step number from the phase's state machine, which increments monotonically during execution and invalidates cached state when operations need to be redone. This step counter enables the frontend to detect when previously completed operations have been invalidated by spiral loop corrections, ensuring UI consistency with the actual generation state.
The step value provides a temporal marker for the phase completion, allowing correlation with intermediate operation events that share the same step number.
Unique identifier for the event type.
A literal string that discriminates between different event types in the AutoBE system. This field enables TypeScript's discriminated union feature, allowing type-safe event handling through switch statements or conditional checks.
Examples: "analyzeWrite", "databaseSchema", "interfaceOperation", "testScenario"
Event fired when the Multi-lingual agent begins the multi-lingual support process.
The multi-lingual support process that begins with this event will proceed through database, interface, test, and realize phases to generate multi-lingual support for the application.
Author
sunrabbit123