User-centered observability: load testing, real user monitoring & synthetics
summary
Shows a concrete stack: embed an open‑source JavaScript agent for real user monitoring to catch a rising “order not visible enough” checkout error, add it to a watchlist, then run a Grafana Cloud Synthetics multi‑HTTP check (public preview, k6‑based) that extracts a product ID via JSONPath and asserts recommendations exist. When that sporadic failure appears, run a k6 cloud performance test bundling browser and protocol scripts; results show no failures but jagged request rate and non‑flat response times and spiky web vitals, prompting further root‑cause tracing.
AI-written summary · about
chapters (11)
- 0:00Introductions
- 1:00Why the end user experience is crucial!
- 2:17Performance golden rule
- 3:39Where to start? Real user monitoring (RUM)
- 4:35Why do you need synthetics if you already have RUM?
- 5:31Why you need performance testing as an engineer
- 6:18The Swiss cheese model
- 7:23Demo - Frontend Observability
- 11:45Demo - Grafana Cloud Synthetics
- 16:03Grafana Cloud k6: Performance testing
- 18:31Summary
related talks
56mFull-stack observability and application monitoring with
26mTo see what others can not... You must climb the mountain, A Introduction to Load Testing
29mHow to Understand End-User Experience with Synthetic Tests, Load Tests and Real User Monitoring
32mCustomer Centric Monitoring
52mTalk with the Experts
32mObservability for Frontend Engineers
28mWhere DevOps Meets QA: An Observability & Synthetic Testing Success Story
53mObservable Web Applications