Skip to main content
George Khananaev
Tag archive

#rate-limiting

2 posts

Read Vercel Controller: a Node.js microservice for Vercel cache and deployments
Technical Deep-Dive
Moon Holidays

Vercel Controller: a Node.js microservice for Vercel cache and deployments

Express 4 service with async job queue, smart deduplication, LRU cache, Helmet.js security, and 129 tests — called by Travel Panel whenever content changes

Vercel Controller wraps the Vercel API behind a stable interface for every service in the Moon Holidays platform. Express 4 + Helmet.js, async job queue with smart deduplication, LRU response cache with 88% hit rate, bearer token auth with timing-safe comparison, 129 tests, and auto-healing Docker. Travel Panel calls it whenever content changes to invalidate CloudFront and redeploy dependent frontends.

nodejsexpressverceldevopsasyncrate-limitingsecuritymicroservice
Jan 20265 min
Read Travel Panel WebSocket Server: real-time at uWebSockets.js speed
FeaturedTechnical Deep-Dive
Moon Holidays

Travel Panel WebSocket Server: real-time at uWebSockets.js speed

High-performance WebSocket gateway for notifications, chat, and live state, running on port 8965 since August 2025

A high-throughput WebSocket server powering every live update across the Moon Holidays platform. Built on uWebSockets.js for raw performance, with MongoDB for persistence, Redis for pub/sub across pods, Firebase Auth for handshakes, and rate limits of 1000 msg/min in dev and 120 msg/min in production.

websocketuwebsocketsnodejstypescriptmongodbredisrealtimerate-limiting
Aug 2025 — Present3 min