Asynchronous programming might sound very compelling, but it is very difficult to master.

There are several pitfalls that you need to be aware of, such as the following:

• Race condition: If two or more threads of code modify the same data, the order in which they get executed can affect the final value. This race can lead to data being in an undetermined state.
• Starvation: Indefinite waiting by one thread due to other threads coming in.
• Deadlock: If a thread is waiting for a resource that another thread has locked, and vice versa at the same time, then both threads are stuck in a deadlock.
• Debugging challenge: It is very hard to reproduce a bug in asynchronous code due to the non-deterministic timing of a multithreaded program.
• Order preservation: There might be dependencies between sections of code that might not be observed when the execution order varies.

In Python, it might be impossible to completely avoid such pitfalls, but we can follow some best practices to eliminate them for most practical purposes. They will be covered in the Celery best practices section.