Decoding Synthetic Monitoring: A Journey from E2E UI Tests to Service-Level Probes
summary
Team replaces flaky Selenium E2E checks with API-level collections (Postman/Bruno) executed as probes, cutting run times from minutes (3–4 min) to ~10–12 seconds in real cases and a demo from 30s to 3s. A Dockerized Bruno probe implements the multi-target exporter pattern (think bespoke Prometheus/blackbox style checks) called by an Open Telemetry collector into Honeycom, enabling higher-frequency probing, Prometheus blackbox exporter compatibility, fewer false alarms and 5–10 minute triage to root cause.
AI-written summary · about
chapters (11)
- 0:00Synthetic monitoring journey
- 3:11UI tests to production monitoring
- 8:11Synthetic monitoring principles
- 12:02UI synthetic monitoring challenges
- 16:35Transitioning to service probes
- 21:11API probes framework
- 25:25Demo: UI to API probes
- 31:46Unified observability benefits
- 35:58Performance and alert gains
- 38:05New approach advantages
- 39:11Q&A
related talks
40mDecoding Synthetic Monitoring: A Journey from E2E UI Tests to Service-Level Probes
30mSynthetic Monitoring and E2E Testing: 2 Sides of the Same Coin
34mReliability Monitoring: Tips and Tools
32mSynthetic Monitoring and E2E Testing: 2 Sides of the Same Coin
44mSynthesizing Sanity with, and in Spite of, Synthetic Monitoring
38mFreeing up Selenium
37mSelenium In the Enterprise: What went right, what went wrong (so far)
36mDeep Dive: Prometheus