Catching production errors before our customers
summary
A real incident shows a SOAP integration failing only in production because a null byte slips into the response; reproduction on local machine loses the null and hides the bug. The workflow uses tcpdump and Wireshark to capture packets, then adds Sentry SDK (pip install sentry-sdk, five lines for Django) to capture exceptions with user, params, traceback and grouped issues. Sentry runs open source or SaaS (free tier) and notifies teams so errors get fixed before customers call.
AI-written summary · about
chapters (17)
- 0:00Introduction
- 0:42What happened next
- 1:04What happens in production
- 1:21What happened in my previous job
- 1:47Sometimes the feature was not working
- 2:05Debugging
- 3:04Testing
- 3:25TCP dump
- 4:16Missing data
- 4:45What is Sentry
- 5:33Django integration
- 6:06Python integration
- 7:22Demo
- 7:45Sensory Interface
- 8:36Projectors
- 9:25Incident Flow
- 10:44Questions Answers
related talks
23mObserving your applications with Sentry and Prometheus
12mSolve 100% of your errors
38mSentry as a way not to be afraid
14mSpeed up fixing app issues using Sentry
08mMonitoring Errors and Slowdowns Across JS Applications
41mTroubleshooting Spring Boot applications with Sentry
49mObservability in Practice: How to Become Proactive with Bugs
33mFrom “its broken” to “fix is out”