Continuous Monitoring + Continuous Testing in DevOps delivery models
summary
Built an in-house tool called ImageVision that runs visual monitoring against live time-series plots and UI regions, integrates with Protractor/Playwright via a single API call, and executes image-based interactions and comparisons in production-like streams. The stack pairs TypeScript front-end TDK with Python CV engines and three modules Image Grab/Image Compare/Image Interact; it handles 6–8 hour runs over hundreds of images, flags timestamped anomalies, and reduced pixel-match false fails by ~80% in practice.
AI-written summary · about
chapters (9)
- 0:00Start
- 1:39Applications and Releases
- 2:24Why We Need Continuous Monitoring
- 3:26Single Track Anomalies
- 6:18Overview of Testing versus Monitoring
- 7:42Monitoring by Plotting
- 13:11Visual Automation and Live Stream Monitor
- 13:31Image Comparison Module
- 17:42Visual Automation
related talks
12mIntegrating visual testing into the build pipeline
20mVisual testing with Visual Regression Tracker
27mVisual UI Testing – Testing the Last Mile
32mContinuous Visual Integration for Rails
34mENABLING CONTINUOUS DEPLOYMENT OF PIXEL-PERFECT USER INTERFACES
42mHow Visual AI Makes Testing Easy
08m“Visual Testing”
21mVisual Regression Testing: From a Tool to a Process