Caching everywhere
summary
Boozer uses Django cache framework with Redis/Memcached backends, Python’s lru_cache and cached_property for memoization, and django-cacheback to serve stale values while refreshing in background; nginx and CDN leverage surrogate headers like s-maxage to avoid origin hits. Practical fixes include a Django cache key/thread-safety bug and reuse of pooled Redis clients. Tradeoffs: stale-but-fast pricing/stock updates vs strict freshness. These measures materially reduce latency and origin load and enable scale.
AI-written summary · about
chapters (8)
- 0:00Introduction to caching
- 1:21Python caching techniques
- 5:14Memoization and libraries
- 9:13Django caching framework
- 12:44Data staleness and validation
- 16:40Page and global caching
- 21:13Application layers and CDNs
- 24:03Network latency and headers
