Resiliency Testing with Toxiproxy
summary
Shopify builds and runs an open‑source TCP proxy called Toxiproxy that sits between apps and dependencies, exposes an HTTP API, and has Ruby and Go clients so tests can inject latency, blackholing and connection rejects. It runs in dev and CI to turn incident fixes into automated assertions, prevents regressions with hundreds of Toxiproxy tests, and complements production game days like weekly automated flash‑sale drills. The tool reliably reproduces network failures and ships on GitHub.
AI-written summary · about
chapters (12)
- 0:00Start
- 2:35Reasoning about failure is hard.
- 2:54Constantly changing.
- 3:49Root cause?
- 10:00Resiliency is a product concern
- 11:23Chaos Engineering
- 13:42Reactive Testing
- 15:14assert application_can_boot end
- 16:59A few hundred Toxiproxy Tests
- 18:10Proactive Testing
- 20:22What's next?
- 22:30Go read the readme for more information!
related talks
28mBuilding and testing resilient applications
16mHijacking Service Discovery to Simulate Dependency Degradation
25mResilience testing in modernization projects
40mStrategic Business Functionality Testing: Accelerating Development Through Mocking
48mClimbing the Testing Pyramid: From Real Service to Interface Mocks in Go
27mEffectively Testing Services
40mResilience Testing - Failure is an Option
39mTesting Level Hard: Distributed Apps