Skip to content

Blog

Notes from the edge

Engineering deep-dives, release notes, and the decisions behind Vega — written by the people who ship it.

Latest launch edge caching

Vega 2.0: edge caching is GA

After six months in beta and 41 billion requests served, edge caching is generally available on every Vega plan. What shipped, what changed since the beta, and how to turn it on.

Priya Raghavan

Co-founder & CTO ·

Read post

engineering

How we cut p99 latency to 42ms with edge caching

A deep-dive into the cache layer behind Vega 2.0: stale-while-revalidate at all 24 regions, single-flight revalidation, and the tag-based invalidation model that made it safe to turn on.

Mara Chen

observability

Tracing distributed requests without the grief

Most tracing rollouts die in the collector config. Vega traces every request end-to-end by default — no agents, no sampling YAML, no separate backend to keep alive at 2am.

Tomás Rivera

engineering

Why we built our own deploy pipeline

Conventional CI treats a deploy as an event. Ours had to be a keystroke. The reasoning behind the 1.2-second build, the tradeoffs we accepted, and what we would do differently.

Sam Okafor