Execution Logging¶

The Admin Panel integrates with ToolRegistry's execution logging feature to provide detailed insights into tool usage.

Enabling Execution Logs¶

from toolregistry import ToolRegistry

registry = ToolRegistry()

# Enable logging with custom buffer size
log = registry.enable_logging(max_entries=1000)

# Register and use tools...
@registry.register
def calculator_add(a: int, b: int) -> int:
    return a + b

# Enable admin panel to view logs
info = registry.enable_admin()

Log Entry Structure¶

Each execution log entry contains:

Field	Type	Description
`id`	`str`	Unique identifier (UUID)
`tool_name`	`str`	Name of the executed tool
`timestamp`	`datetime`	When the execution occurred
`status`	`ExecutionStatus`	`success`, `error`, `timeout`, or `disabled`
`duration_ms`	`float`	Execution duration in milliseconds
`arguments`	`dict`	Input arguments passed to the tool
`result`	`Any`	Execution result (for successful executions)
`error`	`str \\| None`	Error message (for failed executions)
`exception_type`	`str \\| None`	Qualified exception class name, e.g. `"ValueError"`
`traceback`	`str \\| None`	Formatted traceback string from the exception
`metadata`	`dict`	Additional metadata

Querying Logs Programmatically¶

# Get the execution log instance
log = registry.get_execution_log()

if log:
    # Get recent entries
    entries = log.get_entries(limit=10)

    # Filter by tool name
    calc_entries = log.get_entries(tool_name="calculator_add")

    # Filter by status
    from toolregistry.admin import ExecutionStatus
    errors = log.get_entries(status=ExecutionStatus.ERROR)

    # Get statistics
    stats = log.get_stats()
    print(f"Total executions: {stats['total_entries']}")
    print(f"Average duration: {stats['avg_duration_ms']:.2f}ms")