Containerized Time Travel: Replicating Production Performance
summary
A bank serving 200M customers across 150+ countries recreates real production load by recording traces with OpenTelemetry, alerting with Grafana, then using Speed Scale record+replay to capture a problematic user flow, sanitize PII, and replay it in dev with mocked downstream services. Teams reproduce issues in ~2 hours instead of ~24, cut a hotspot from ~180 ms to ~20 ms, and report a 30% ops cost reduction and 25% transaction efficiency gain. This is containerized time travel using sanitized snapshots.
AI-written summary · about
chapters (18)
- 0:00Containerized time travel intro
- 2:35Production outages impact
- 5:11Dev vs prod paradox
- 7:01Why reproduce issues
- 9:10Microservices challenges
- 12:42Current approaches limits
- 13:44Safe workload visibility
- 14:25Containerized time travel concept
- 16:58Four-step issue solving
- 17:59Banking app demo walkthrough
- 19:27Step 1: Monitoring issues
- 20:32Step 2 traffic capture with Speedscale
- 22:18Step 3: Traffic sanitization
- 25:53Step 4: Intelligent replay
- 27:18Case study results
- 28:22Key benefits
- 29:50Call to action
- 30:38Q&A
related talks
44mPerformance Testing with a "Big Data Tool"
32mCustomer Centric Observability: How Intuit Reduced Time To Detect Customer Impact From 30+ Minutes To Under 3 Minutes.
31mUnlocking the Power of OpenTelemetry: Enhancing Design, Development...
42mReplay Production Traffic Locally: Instantly Spot & Fix Breaking Changes
29mFrom Real Traffic to Reliable Integration Tests with AI
56mGetting all the 99.99(9) you always wanted
46mTroubleshooting Distributed Workloads ...
30mArchitecting High-Density Performance Testing with Container-Based Simulation