The Next Evolution of Chaos Engineering Tooling: Introducing Latency Squeeze Injection (LSI)
summary
Claims a new method called Latency Squeeze Injection (LSI) implemented in the PRECog tool that slowly injects latency to map the SLO safety margin rather than causing 100% failures. Uses AWS FIS/SSM agents, Lambda layers and Linux tc to queue packets; pilot on a high‑TPS login flow shows a 200 ms SLO actually breaks around the 170 ms edge, prompting moved circuit breakers, tighter timeouts and connection validation and averting likely Sev3 incidents. Runs daily in QA, aiming for production.
AI-written summary · about
chapters (9)
- 0:00Latency Squeeze Injection intro
- 1:06What is Chaos Engineering
- 4:19Traditional chaos limits
- 13:06Introducing Latency Squeeze Injection concept
- 19:26How LSI works
- 21:37Pilot at Capital One
- 24:49LSI benefits and scale
- 26:57Key takeaways
- 28:21Q&A
related talks
31mPlanning for Failure using Chaos Engineering
05mReducing Trauma in Production with SLOs and Chaos Engineering
29mReducing Trauma in Production
35mFrom Monitoring to Observability: Left Shift your SLOs with Chaos
44mApplying principles of chaos engineering to serverless
38mScientific and Safe Chaos Engineering
55mIncident Management at Netflix Velocity
28mBuilding Safety, Security and Resilience through Chaos