Python Concurrency and Performance

Overview

Correct concurrency starts with matching the model to the workload, not the developer's preference. This skill encodes defaults for model selection, cancellation/deadline behavior, and lifecycle safety—prioritizing explicit control over implicit magic.

Treat these recommendations as preferred defaults. When project constraints demand deviation, call out tradeoffs and compensating controls.

When to Use

• Selecting between asyncio, threading, multiprocessing, or concurrent.futures
• Propagating deadlines or cancellation through async call chains
• Bounding fan-out, backpressure, or semaphore-guarded concurrency
• Diagnosing race conditions, deadlocks, or priority inversion
• Profiling throughput bottlenecks before and after optimization
• Verifying no task or thread leaks on shutdown or lifecycle transitions