dify_admin/docs/api/operation_logger.md

Operation Logger API Documentation

Overview

The OperationLogger class provides functionality for recording and managing operation logs, including:

• Logging API operations
• Querying operation logs with filters
• Cleaning up old logs

Class Methods

log_operation(user_id, operation_type, endpoint, parameters=None, status="SUCCESS")

Records an API operation in the database.

Parameters:

• user_id (int): ID of the user performing the operation
• operation_type (str): Type of operation (e.g., "CREATE", "UPDATE")
• endpoint (str): API endpoint accessed
• parameters (str, optional): Operation parameters in string format
• status (str): Operation status (default: "SUCCESS")

Returns:

• int: ID of the created log record

Example:

logger.log_operation(123, "LOGIN", "/api/auth/login", "username=test", "SUCCESS")

get_operations(user_id=None, start_time=None, end_time=None, limit=100)

Queries operation logs with optional filters.

Parameters:

• user_id (int, optional): Filter by user ID
• start_time (datetime, optional): Earliest log time
• end_time (datetime, optional): Latest log time
• limit (int): Maximum number of logs to return (default: 100)

Returns:

• List of log records with fields:
  • id
  • user_id
  • operation_type
  • endpoint
  • parameters
  • status
  • created_at

clean_old_logs(days=30)

Deletes logs older than specified number of days.

Parameters:

• days (int): Delete logs older than this many days (default: 30)

Returns:

• int: Number of logs deleted

Error Handling

• Raises OperationLogError for all failures
• Detailed error logging through module logger

Best Practices

• Log all critical operations for audit trail
• Regularly clean old logs to maintain performance
• Use appropriate status values ("SUCCESS", "FAILED", etc.)
• Include relevant parameters for debugging