Automated Troubleshooting of Live Site Issues
summary
PayPal builds an Auto Troubleshooting Platform that ingests tickets and alerts from a triage server, enriches them via an associative search and utility scraper, runs a rules-backed workflow finder to call over ten pluggable fetchers, and combines results in an assimilation service logged to a portal. The system targets a 2000+ application stack, returns automated diagnoses in minutes, reduces MTTR and manual effort, and supports adding workflows and language-agnostic fetchers.
AI-written summary · about
chapters (11)
- 0:00Introduction
- 0:18About Me
- 2:08Agenda
- 3:27Team Overview
- 3:59Troubleshooting Challenges
- 7:00Troubleshooting Manual
- 8:54Implicit Functional Requirements
- 10:50Architecture
- 17:38Workflow
- 18:26Features
- 19:28Benefits
related talks
46mSupercharging PayPal’s Application Development
20mAI Powered Root Cause Analysis for Production Alerts
14mTracing: Frontend Issues With Backend Solutions
18mProduction software keeps breaking and it will only get worse
33mFrom “its broken” to “fix is out”
09mAutomated Application Security Testing
30mBeyond the Dashboard: Build End-to-End O11y
49mAerospike Management APIs and Operations at Scale