Logging¶

AI Services in AIDF support two complementary log types:

1. Verbose Logs (Structured)¶

Structured, high-level execution entries stored in the database (AI_SERVICE_EXECUTION_LOGS) and visible in the pipeline UI. Each entry includes:

Timestamps for start/end of each step
Total duration
Number of input/output records
Status (success/failure)
Error messages, if any

Useful for tracking service performance, auditing, and reviewing processed data summaries.

2. Console Logs (Real-time)¶

Real-time messages printed to terminal/log stream for development and debugging:

logger.info("TTS model loaded successfully")        # High-level milestones
logger.debug(f"Intermediate value: {variable}")      # Debug variables
logger.error(f"Exception: {str(e)}")                 # Error traces

Rule

Every function must be wrapped in a try/except block. Log exceptions with logger.error() before raising or returning an error response.