Python at Scale: Architecting for Speed: Beyond Simple Fixes: A Guide to Building Truly Performant, Data-Intensive Systems - Softcover

F. Main, Kevin

 
9798273157453: Python at Scale: Architecting for Speed: Beyond Simple Fixes: A Guide to Building Truly Performant, Data-Intensive Systems

Synopsis

Let's be honest: your Python application is slow.
You’ve already done the "simple fixes." You've added a database index, optimized a loop, and maybe even pip install-ed a new library that promised magic. And yet, your P99 latency is climbing, your pandas script just OOM-killed another server, and your users are starting to notice.
Here’s the hard truth: your problem isn't your code; it's your architecture.
Python at Scale: Architecting for Speed is not another book of clever "hacks." It's a hands-on guide to thinking like a systems architect, for Python developers who have hit the performance wall. We're going "beyond simple fixes" to build truly performant, data-intensive systems that are fast by design.
In this book, we'll stop tweaking and start re-architecting. You will learn to:

  • Identify the Real Villain: Stop guessing. Learn to use the right profiler (cProfile, Py-Spy, Scalene) to find your actual bottleneck—whether it's I/O wait, CPU (the GIL), or memory.
  • Conquer Concurrency: Go beyond a simple await. Master the asyncio event loop, understand its "blocking" traps, and learn to build systems that can actually handle thousands of concurrent requests.
  • Shatter the GIL: Stop fighting the Global Interpreter Lock. Learn to identify CPU-bound problems and use multiprocessing to shatter the "one-core" barrier and use your hardware's full power.
  • Build an Impenetrable Data Layer: Your database is the bottleneck. Learn to protect it with advanced caching (Redis), denormalization, and the bulletproof "Outbox Pattern" to decouple your writes.
  • Decouple Everything: Your app shouldn't do things "right now." Learn to use message queues like RabbitMQ and Kafka to build resilient, asynchronous, and blazing-fast services.
  • Kill Your pandas Script (Before It Kills Your Server): That in-memory script is a time bomb. We'll show you how to move to out-of-core, parallel tools like Dask and Polars to process 500GB of data as easily as 5MB.
  • Achieve True Observability: Learn to deploy and monitor with confidence. We'll cover everything from Docker and Kubernetes to the metrics that actually matter (hint: it's not "average latency").
This book is your guide to moving from a developer who fights fires to an architect who prevents them. Stop "fixing" your code and start building systems that scale.

"synopsis" may belong to another edition of this title.