Event loop management

Enables access to the current event loop for scheduling and running async tasks from synchronous code

Legacy async code: manually running a coroutine from a synchronous entry point, Library integration: bridging sync and async code boundaries, Testing: obtaining a loop to run async test fixtures

asyncio applications, legacy Python async code, test suites for async functions

loop = asyncio.get_event_loop()

asyncio.get_event_loop()

loop: asyncio.AbstractEventLoop object

• loop.run_until_complete()
• loop.run_forever()
• asyncio.run()
• asyncio.new_event_loop()

• asyncio.run(coro): preferred replacement in Python 3.7+ that creates a new loop and runs the coroutine to completion
• asyncio.new_event_loop(): creates a fresh loop without checking for an existing one

Calling get_event_loop() when no loop exists in a non-main thread raises DeprecationWarning in 3.10+ and will error in future — misconception that it always creates a loop; consequence: runtime errors in thread workers, Calling loop.run_until_complete() inside an already-running loop — misconception that nesting is allowed; consequence: RuntimeError about nested event loops, Using get_event_loop() instead of asyncio.run() in simple scripts — misconception that manual loop management is required; consequence: verbose boilerplate and missed automatic cleanup

asyncio.run(): creates and manages a fresh loop for the entire coroutine lifecycle, asyncio.new_event_loop(): explicitly creates a new loop without referencing the current one, asyncio.get_running_loop(): returns the currently running loop or raises if none is running

Coming from JavaScript: may expect get_event_loop() to start processing like the Node.js event loop — Python's loop must be explicitly run with run_until_complete() or run_forever()

• asyncio.get_running_loop()
• asyncio.new_event_loop()
• asyncio.run()
• loop.run_until_complete()

Deprecated in Python 3.10+ when called where no running loop exists; prefer asyncio.run() for new code. In Python 3.12+, calling from a thread with no running loop raises an exception rather than creating one. Creating a loop without running it can leak resources if not closed.

Provides a handle to the event loop for manual coroutine execution, enabling integration of async code into synchronous contexts at the cost of explicit lifecycle management

Like asking the front desk for the current event venue — it gives you the existing one or sets one up, but in newer buildings they prefer you book through the official concierge (asyncio.run).

Since Python 3.10, asyncio.get_running_loop() should be used inside coroutines and callbacks. The deprecation path leads to removal of implicit loop creation behavior.

asyncio.run()