Skip to main content

Concurrency with `threading`

Standard Library

threading Module Guide

Threads shine for I/O-bound workloads: API calls, file operations, and background tasks. Learn the primitives you need to stay safe.

Spin up threadsCoordinate safely

Basics​

import threading

def worker(order_id):
    process(order_id)

threads = [
    threading.Thread(target=worker, args=(oid,))
    for oid in order_ids
]
for t in threads:
    t.start()
for t in threads:
    t.join()

Synchronization primitives​

PrimitivePurpose
Lock / RLockMutual exclusion around shared state
EventSignal between threads
SemaphoreLimit concurrency
ConditionWait for state changes
QueueThread-safe producer/consumer communication

Prefer queue.Queue for coordinating work—no explicit locking required.

When not to use threads​

  • CPU-bound tasks are limited by the GIL; use multiprocessing or native extensions instead.
  • Threads add complexity—profile before introducing them.

Next up in your learning path

multiprocessing

Parallelize CPU-bound work across processes.

Parallel

asyncio (future cluster)

Consider async for large numbers of concurrent network tasks.

Preview

Performance pillar

Track metrics and profiling when using threads.

Future