Talks - Joongi Kim: Improving debuggability of complex asyncio applications

Improve debuggability of complex asyncio applications by understanding its limitations, using the right libraries, and leveraging tools like time module, _syncio library, and STEP debugger to inspect and control execution.

Key takeaways

Improving debuggability of asyncio applications starts by understanding the basics of asyncio and its limitations.
Coroutines and tasks in asyncio are inherently lightweight and should be used to handle I/O-bound tasks, not CPU-intensive tasks.
Using libraries like trio and curio can simplify development and improve performance.
async/await syntax can help write more readable and synchronous-like code, but should be used with caution due to the overhead of switching between the event loop and the CPU.
To debug asyncio applications, use tools like the time module, _syncio library, and STEP debugger to inspect and control the execution of coroutines and tasks.
When debugging asyncio applications, use a narrowed exception message and inspect the stack trace to identify the problematic code.
If using libraries like async_timeout or aiosmtpd, ensure that they are configured correctly and do not interfere with the event loop.
When using pytest in a Jupyter notebook, use the pytest-asyncio library to properly run and test asyncio tests.
To improve the performance of asyncio applications, use a profiler like line_profiler or snakeviz to identify performance bottlenecks and optimize the code.
To reduce the risk of memory leaks in asyncio applications, use tools like mypy and pytest to identify potential issues before they become a problem.

Talks - Joongi Kim: Improving debuggability of complex asyncio applications

More talks