It Was Fast on My Machine: What Production Reveals About Web Performance
summary
Local-fast 120 ms responses routinely jump to 2.8 s in production because of real-world layers: DNS, CDN misconfiguration, differing cache hit vs cache miss behaviour, wrong HTTP cache-control headers, pod CPU/memory limits and database misses. The talk shows concrete fixes: use stale-while-revalidate, set correct TTLs per layer, add tracing and logs in OpenShift to find the slow layer, and run load tests with autocannon to validate performance under real traffic.
AI-written summary · about
chapters (13)
- 0:00Performance differences local vs prod
- 1:51Why production slows down
- 3:59Request to database journey
- 6:24User-centric performance metrics
- 7:28Cache layers explained
- 10:17Cache misses in production
- 12:26HTTP headers and caching
- 13:33CDN myths and pitfalls
- 14:58Prod runtimes vs local limits
- 16:47Measuring production performance
- 18:52Mindset shift to production
- 20:58Production checklist
- 22:27Q&A
