Peter Paul Bakker: The Power of Automated Resilience Tests
summary
A tiny bank demo runs automated resilience tests in CI using k6 + WireMock + test-event-scheduler to inject delays (2s and 10s) and restarts while collecting JFR and HTTP-client metrics. Default Apache HTTP client settings cause threads and memory to spike and requests to hang; changing one line to set connection request timeout to 100ms cuts 60s hangs, then sizing pools to 60/30 connections removes backlog and restores responses, though a 10s backend still needs a 2–3s call timeout or fallback.
AI-written summary · about
chapters (16)
- 0:00Introduction
- 1:31Speaker background
- 4:01Resilience concepts
- 7:10Resilience strategies
- 12:54Tiny Bank story
- 15:43Automating with k6 and WireMock
- 17:30First test and results
- 23:16Bottlenecks and memory issues
- 25:48Fixing timeouts and pools
- 28:46Second test improvements
- 30:38Extreme delays and fallbacks
- 32:20Resilience testing blocks
- 34:04Lessons learned
- 35:51Simplicity versus resilience
- 37:25Preventing failures
- 38:26Q&A
related talks
25mResilience testing in modernization projects
42mPlanning To Fail
01h 01mBreaking your REST APIs to make them stronger!!
40mResilience Testing - Failure is an Option
42mHow to stay alive even when others go down:...
28mRevealing Resilience Vulnerabilities in Spring Boot Architectures
21mCrunching Millions of Requests for Breakfast: A Tale of Performance Testing Prowess
21mEmbracing Resilience in CD Pipelines