Caching the uncacheable: delivering personalized experiences without sacrificing performance
summary
Shows how Varnish turns personalization into cacheable work by using Varnish VCL to trim or keep specific cookies, create language-based cache variants, and serve synthetic JSON from Redis-backed session data so 99% cacheable pages don’t hit origin for the 1% of stateful bits. Demonstrates ESI subrequests for stitched placeholders, cookie VMOD and digest modules, on-edge JWT validation and signature checking, and real-world throughput numbers (1.3 Tbps single machine) and TTL strategies.
AI-written summary · about
chapters (14)
- 0:00Caching challenges
- 4:17Slow sites lose money
- 11:02Reverse proxies protect apps
- 13:29HTTP caching mechanisms
- 20:11Cookies break caching
- 23:42VCL explained
- 25:51Custom caching with VCL
- 34:19Cache variations with cookies
- 39:45Partial caching with ESI
- 40:47Compute on the edge with VCL
- 42:12JWT validation at edge
- 47:08Redis integration
- 55:39Future of edge compute
- 57:49Q&A
