Push-Button Reliability Testing for Cloud-Backed Applications with Rainmaker
summary
A push-button SDK called Rainmaker injects faults at the REST-layer fault injection using an HTTP proxy and automatic fault policies so existing test suites run without new oracles or workloads. It detects real issues — 73 bugs (55 confirmed, 51 fixed) across 11 cloud-backed apps — including semantic violations, unrelated exceptions, and state divergence. Rainmaker generates optimized test plans and reduces exhaustive runtime from ~580 days to about 56 hours while keeping false alarms low.
AI-written summary · about
chapters (13)
- 0:00Intro
- 0:38The emerging cloud-based programming model
- 1:49Dark side: new reliability challenges
- 3:37Does retry solve all the problems?
- 6:47Design goals of Rainmaker
- 7:45A taxonomy of error-handling bugs
- 8:45Silent semantic violations
- 9:13State divergence
- 10:24Rainmaker's fault injection policies
- 11:21Rainmaker performs HTTP layer injection
- 11:58Rainmaker reuses existing test oracles
- 14:21Generating test plans
- 15:21Finding new bugs
related talks
27mQuality on Cloud 9: 9 Aspects of Delivering a Great Product on Cloud
57mLast- und Performancetests in der Cloud?!
01h 08mPerformance testing in the cloud
16mWho Watches the Watchers? On the Reliability of Softwarizing Cloud Application Management
40mResilience Testing - Failure is an Option
28mPush button deployments in minutes not months 🧭
12mExponential Load Testing...
33mPerformance Testing in Agile and DevOps