The Benediction blog
Engineering deep dives and product notes from the team. Subscribe via RSS.
Edge caching, explained: from cache keys to stale-if-error
How our cache decides what to store, what to invalidate, and what to serve when an origin goes quiet. Includes a worked example you can reproduce.
Serverless at the edge: architecture we wish we'd used sooner
Workers are not a smaller Lambda. The mental models that transfer — and the ones that don't — learned from serving 200B requests a year.
TLS 1.3 and HTTP/3: what's different at the edge
Why we ship TLS 1.3 everywhere, what enabling HTTP/3 changed for real users, and how the protocol stack fits together.
Tuning the WAF without breaking your app
A practical playbook for rolling out managed rules in simulation mode, triaging matches, and shipping custom rules with confidence.
Building with Edge KV: from feature flags to cart state
Three production patterns for global key-value storage, plus the consistency model we use and the mistakes we made early on.
Zero-downtime deploys with instant rollbacks
How versioned Worker deploys work under the hood, and why a one-second rollback changes how you think about shipping.