It's not how good your app is, it's how good you want it to be
summary
Claim: visibility wins. Concrete wins: Travis CI runs thousands of test runs daily (8,500+ and 1.5M total), uses ActiveSupport Notifications and a custom Travis Instrumentation layer to emit events, ships metrics through statsd / metricsd into Librato and Graphite, and traces non‑HTTP work with NewRelic. Built tooling includes travis admin, an event monitor and travis logs; centralized Papertrail/Haystack/Airbrake capture logs and exceptions so the distributed VM workers and message hub become observable and actionable.
AI-written summary · about
chapters (18)
- 0:00Start
- 0:40Presentation
- 5:08It's Not How Good You Are It's How Good You Want To Be
- 5:48App Is Not a Black Box
- 10:01Basic Architecture
- 12:21Measuring and Monitoring
- 12:33Logging
- 17:11Consolidate All Our Logging
- 18:31Exceptions
- 18:52Exception Notifier
- 21:28Messaging Hub
- 24:25Parts to Metrics
- 24:44Active Support Notifications
- 31:34Login Reporter
- 32:33Visualization
- 32:57Librato
- 36:00Internal Tools
- 37:40Event Monitor
related talks
44mYour App is not a Black Box
12mTravis CI
35mScaling an Open Source CI with a Remote Team, the Story behind Travis CI
20mContinuous integration for mobile
28mDo You Know If Your Service Is Working Properly?...
21mMonitoring Your Apps is not About Devops
39mRunning Ruby Apps
28mKeeping Developers Happy with a Fast CI