Pytest: Борьба с Race Conditions в Async-коде
ТОЛЬКО для async-разработчиков: race conditions, flaky tests, async fixtures — специализация для asyncio.
What you'll learn
Специализированный курс ТОЛЬКО для async-разработчиков: pytest-asyncio, race conditions в PostgreSQL/Redis, flaky tests, resilience patterns. ВАЖНО: Если вы НЕ работаете с async/await ежедневно — этот курс не для вас. 8 часов глубокого погружения в async-тестирование.
⚠️ ONLY for async developers
This course is NOT for sync code. If you DON'T use async/await daily—take 'DB Testing' (Level 4) instead of this course.
Full cycle of flaky test handling
From diagnosis to production fix: reproduce race conditions, analyze causes, fix with proper patterns. Structured approach from simple to complex.
Practical learning cases
Simplified examples of typical production problems: promo codes twice, double charges, Redis failures. Enough to understand patterns, but not full production complexity.
Pure async, no compromises
Entire course — async code only. pytest-asyncio, asyncpg, aioredis. No sync examples. Deep dive into async testing.
Automated detection
Write your own pytest plugin for automatic flaky test detection, CI integration, quarantine of unstable tests, and quality gates.
Testimonials
alumni“Nondeterministic tests have two problems. First, they are useless. Second, they are a virulent infection that can completely ruin your entire test suite.”
Martin Fowler
“Tests should be deterministic. If they aren't, you can't trust them, and if you can't trust them, they're worthless.”
Kent Beck (создатель TDD)
“Flaky tests are one of the most harmful things that can happen to a test suite. They erode confidence and waste enormous amounts of engineering time.”
Google Testing Blog
What's inside
Async fixtures for PostgreSQL and Redis: transactions, rollback, concurrent operations
Race conditions in learning cases: promo code twice, double charge, stale cache—reproduce and fix in simplified examples
Idempotency and locks: asyncio.Lock, Redis Lock, unique operation keys, retry logic testing
Resilience testing: graceful degradation, circuit breakers, health checks on service failures
Custom pytest plugin: automatic flaky test detection, CI integration, quality gates
Prerequisites
Pytest Basics: First Tests
requiredMinimum Level 1 (basic tests). Recommended to complete levels 1-3 for deep understanding.
Go to coursePython async/await: practical experience
requiredREQUIRED: minimum 4 hours of async/await, asyncio.gather, asyncio.Lock practice. Course will be incomprehensible without it. Recommended: 'Real Python: Async IO in Python' or asyncio official docs.
PostgreSQL: basic operations
Experience with PostgreSQL: SELECT, INSERT, UPDATE, transactions. Recommended for understanding DB examples.
Course curriculum
Фундамент: окружение и pytest-asyncio
60 минут
Async фикстуры для PostgreSQL
50 минут
Первая async-гонка: промокод дважды
45 минут
Гонки PostgreSQL: lost update, dirty read
55 минут
Async Redis: фикстуры и Cache Aside паттерн
50 минут
Гонки кеша: stale data problem
50 минут
Идемпотентность: двойное списание
55 минут
Resilience: падение сервисов
55 минут
Автоматизация: pytest plugin для flaky tests
60 минут
FAQ
Do I need to take basic courses before this?
Minimum Level 1 (basic tests). Recommended to complete Levels 1-3 for confidence in fixtures, mocks, pytest-xdist. But main requirement—async/await experience: minimum 4 hours asyncio practice.
I don't work with async. Is this course for me?
NO. This course is ONLY about async testing. Without async/await experience, you won't understand the material. First complete minimum 4 hours of asyncio practice, then come back.
Is this course for me if I work with databases?
Only if you use asyncpg/aiohttp/async DB drivers. If working with psycopg2/requests (sync code)—take 'DB Testing' (Level 4). Same patterns but without async.
What should I do after the course?
Find a flaky test in your project (everyone has them!), reproduce the race, write a fix. Then: add async fixtures, set up observability, write your own pytest plugin for flaky detection.
Related Materials
These materials complement the course but are not part of it. Study at your own pace.
DB Testing with PostgreSQL (Level 4)
If you DON'T work with async—start with this course. PostgreSQL, Alembic, factory_boy, Testcontainers—same patterns but for sync code. Alternative to current course.
Race Conditions in multithreaded code
Threading, threading.Lock, race diagnostics. For those working with multithreading, not async. Was part of the course, extracted to materials.
Time mocking in async
How to test async code with delays without waiting. asyncio.sleep, freezegun, retry logic. Specialized topic, not part of the main course.
Observability in pytest
Metrics, traces and logs for race condition investigation. Not part of the course—separate advanced topic.
Performance testing with pytest-benchmark
Quick start with benchmarks in pytest. Not part of the course—separate discipline.