Ready, Steady, Crash!
summary
Uses artillery via NPM with config: warm‑up 10/s for 60s, ramp 10→25 over 120s, cruise 25 for 1200s, crash 100/s for 30s. Metrics show request latency jumping to >30,000 ms and >90% 500s; isolated culprit as a GraphQL query fetching first 600 systems that drives 100% CPU at ~10 RPS. Tried driver pool tuning, an index (DB hits 14,213 → −2,148) and bigger instance; issue vanishes overnight so root cause stays unclear but troubleshooting steps and tooling are validated.
AI-written summary · about
chapters (17)
- 0:00Start
- 3:42Testing the technical limits of our node application.
- 5:00Choose a Testing framework
- 5:15Reasons
- 6:26Targeted towards continuous integration.
- 7:20Decide on the type of requests.
- 9:58Warm Up Arrival rate of 10 virtual users/second for 60
- 12:25What should the first script look like?
- 13:39Scenario Section
- 15:35Artillery will provide you with a summary report which contains detailed metrics.
- 17:49How do we find the culprit?
- 18:32Application issue?
- 19:28Neo4j Driver?
- 20:52Cypher tuning?
- 22:04Configuration issue with db?
- 23:50Build a personal toolbox to help identify and eradicate future culprits.
- 24:09Able to measure, assess and improve the overall performance of our application
related talks
29mReady, steady, crash!
19mKeeping High-Traffic Systems Running When Things Go Wrong
30mAbove the Clouds: Investigating Systems from 32,000 ft.
36mMake It Fast
35mStress Testing as a Culture
23mHow to edge cache GraphQL APIs
47mResolving performance issues in practice
48mHow to not Crash in Prod, Load Testing Crash Course