Python

sys.monitoring — Execution event monitoring

Added in version 3.12.

참고

sys.monitoring is a namespace within the sys module, not an independent module, so there is no need to import sys.monitoring, simply import sys and then use sys.monitoring.

This namespace provides access to the functions and constants necessary to activate and control event monitoring.

As programs execute, events occur that might be of interest to tools that monitor execution. The sys.monitoring namespace provides means to receive callbacks when events of interest occur.

The monitoring API consists of three components:

Tool identifiers

A tool identifier is an integer and the associated name. Tool identifiers are used to discourage tools from interfering with each other and to allow multiple tools to operate at the same time. Currently tools are completely independent and cannot be used to monitor each other. This restriction may be lifted in the future.

Before registering or activating events, a tool should choose an identifier. Identifiers are integers in the range 0 to 5 inclusive.

Registering and using tools

sys.monitoring.use_tool_id(tool_id: int, name: str, /) → None

Must be called before tool_id can be used. tool_id must be in the range 0 to 5 inclusive. Raises a ValueError if tool_id is in use.

sys.monitoring.clear_tool_id(tool_id: int, /) → None

Unregister all events and callback functions associated with tool_id.

sys.monitoring.free_tool_id(tool_id: int, /) → None

Should be called once a tool no longer requires tool_id. Will call clear_tool_id() before releasing tool_id.

sys.monitoring.get_tool(tool_id: int, /) → str | None

Returns the name of the tool if tool_id is in use, otherwise it returns None. tool_id must be in the range 0 to 5 inclusive.

All IDs are treated the same by the VM with regard to events, but the following IDs are pre-defined to make co-operation of tools easier:

sys.monitoring.DEBUGGER_ID = 0
sys.monitoring.COVERAGE_ID = 1
sys.monitoring.PROFILER_ID = 2
sys.monitoring.OPTIMIZER_ID = 5

Events

The following events are supported:

More events may be added in the future.

These events are attributes of the sys.monitoring.events namespace. Each event is represented as a power-of-2 integer constant. To define a set of events, simply bitwise OR the individual events together. For example, to specify both :monitoring-event:`PY_RETURN` and :monitoring-event:`PY_START` events, use the expression PY_RETURN | PY_START.

Local events

Local events are associated with normal execution of the program and happen at clearly defined locations. All local events can be disabled. The local events are:

Deprecated event

  • BRANCH

The BRANCH event is deprecated in 3.14. Using :monitoring-event:`BRANCH_LEFT` and :monitoring-event:`BRANCH_RIGHT` events will give much better performance as they can be disabled independently.

Ancillary events

Ancillary events can be monitored like other events, but are controlled by another event:

The :monitoring-event:`C_RETURN` and :monitoring-event:`C_RAISE` events are controlled by the :monitoring-event:`CALL` event. :monitoring-event:`C_RETURN` and :monitoring-event:`C_RAISE` events will only be seen if the corresponding :monitoring-event:`CALL` event is being monitored.

Other events

Other events are not necessarily tied to a specific location in the program and cannot be individually disabled.

The other events that can be monitored are:

The STOP_ITERATION event

PEP 380 specifies that a StopIteration exception is raised when returning a value from a generator or coroutine. However, this is a very inefficient way to return a value, so some Python implementations, notably CPython 3.12+, do not raise an exception unless it would be visible to other code.

To allow tools to monitor for real exceptions without slowing down generators and coroutines, the :monitoring-event:`STOP_ITERATION` event is provided. :monitoring-event:`STOP_ITERATION` can be locally disabled, unlike :monitoring-event:`RAISE`.

Note that the :monitoring-event:`STOP_ITERATION` event and the :monitoring-event:`RAISE` event for a StopIteration exception are equivalent, and are treated as interchangeable when generating events. Implementations will favor :monitoring-event:`STOP_ITERATION` for performance reasons, but may generate a :monitoring-event:`RAISE` event with a StopIteration.

Turning events on and off

In order to monitor an event, it must be turned on and a corresponding callback must be registered. Events can be turned on or off by setting the events either globally and/or for a particular code object. An event will trigger only once, even if it is turned on both globally and locally.

Setting events globally

Events can be controlled globally by modifying the set of events being monitored.

sys.monitoring.get_events(tool_id: int, /) → int

Returns the int representing all the active events.

sys.monitoring.set_events(tool_id: int, event_set: int, /) → None

Activates all events which are set in event_set. Raises a ValueError if tool_id is not in use.

No events are active by default.

Per code object events

Events can also be controlled on a per code object basis. The functions defined below which accept a types.CodeType should be prepared to accept a look-alike object from functions which are not defined in Python (see Monitoring C API).

sys.monitoring.get_local_events(tool_id: int, code: CodeType, /) → int

Returns all the local events for code

sys.monitoring.set_local_events(tool_id: int, code: CodeType, event_set: int, /) → None

Activates all the local events for code which are set in event_set. Raises a ValueError if tool_id is not in use.

Disabling events

sys.monitoring.DISABLE

A special value that can be returned from a callback function to disable events for the current code location.

Local events can be disabled for a specific code location by returning sys.monitoring.DISABLE from a callback function. This does not change which events are set, or any other code locations for the same event.

Disabling events for specific locations is very important for high performance monitoring. For example, a program can be run under a debugger with no overhead if the debugger disables all monitoring except for a few breakpoints.

sys.monitoring.restart_events() → None

Enable all the events that were disabled by sys.monitoring.DISABLE for all tools.

Registering callback functions

sys.monitoring.register_callback(tool_id: int, event: int, func: Callable | None, /) → Callable | None

Registers the callable func for the event with the given tool_id

If another callback was registered for the given tool_id and event, it is unregistered and returned. Otherwise register_callback() returns None.

Functions can be unregistered by calling sys.monitoring.register_callback(tool_id, event, None).

Callback functions can be registered and unregistered at any time.

Callbacks are called only once regardless if the event is turned on both globally and locally. As such, if an event could be turned on for both global and local events by your code then the callback needs to be written to handle either trigger.

Callback function arguments

sys.monitoring.MISSING

A special value that is passed to a callback function to indicate that there are no arguments to the call.

When an active event occurs, the registered callback function is called. Callback functions returning an object other than DISABLE will have no effect. Different events will provide the callback function with different arguments, as follows:

분실물 보관소