Data Driven Cache Invalidation

on YouTube ↗52m

" summary

Shows how to invalidate HTTP caches from the database by sending Varnish purge requests from Postgres. Uses triggers + PL/Python for a simple synchronous purge (blocking, brittle across multiple writers), then replaces it with an async pipeline: insert events into pgq (SkyTools), run a consumer that calls Varnish admin to purge URLs, and optionally tag cached responses with custom headers. Explains tradeoffs, performance, retry behavior and when per-URL vs broader purges make sense.

AI-written summary · about

" chapters (7)

1. 0:00Introduction and architecture
2. 4:16Application demo setup
3. 10:48Adding varnish and caching
4. 16:32Data-driven invalidation
5. 20:55Trigger-based purging
6. 28:13Asynchronous queueing with PGQ
7. 42:54Advanced strategies and Q&A

" related talks

• 58mCaching the uncacheable: delivering personalized experiences without sacrificing performanceThijs Feryn · Copenhagen Developers Festival 2023
• 28mCaching the uncacheable in VarnishThijs Feryn · SREday London 2024
• 37mVarnish tips & tricksP. Pasich · 4Developers 2015
• 42mCaching for CashKent C. Dodds · JavaScript fwdays’23
• 46mVarnish for PHP developersMattias Geniar · Laracon EU 2016
• 38mThe power of cache in a slow worldMichael May · RailsConf 2015
• 27mCaching for CashKent C. Dodds · UtahJS Conf 2023
• 01h 00mDrupal 8 Caching: A Developer’s GuidePeter Sawczynec · MidCamp 2017