VeloCMS
BlogDocsChangelog
Subscribe
Blogengineering

How We Scaled to 1M Requests per Second

The infrastructure decisions, caching strategies, and architectural patterns that got us from 10K to 1M req/s.

V

VeloCMS Team

April 12, 2026 · 1 min read

Share
How We Scaled to 1M Requests per Second

On this page

Where we started

A single Node.js server behind an Nginx reverse proxy. It worked until it didn't — the first Black Friday showed us exactly where the ceiling was.

The caching layer

Redis for session state, Cloudflare for edge caching, and a custom invalidation pipeline that could flush specific cache keys in under 50ms globally.

The database strategy

Read replicas for the hot path, write-through cache for mutations, and a separate analytics pipeline that never touches the production database.

engineering

On this page

Stay updated

Weekly engineering insights.

VeloCMS
PrivacyTermsRSS
© 2026 VeloCMS