Case Study: Glide scales a 200GB in-memory database and preserves real-time performance with Redis Cloud (Redis Labs)

Glide.me Leverages Redis Cloud to Scale Their 200G In-Memory Database

Glide, a mobile video instant-messaging app serving millions of users, runs on AWS and depends on Redis Cloud for sessions, transient data and caching. They needed persistent in-memory storage that would survive autoscaling and support real-time reminders, live analytics and high-volume PubNub traffic — and they faced a severe client-side bug that inflated session data into a near-200GB "memory leak."

With Redis Cloud, Glide moved unread/missed counts and message lists into Redis (offloading MongoDB), incremented counters for real-time analytics and PubNub metrics, and isolated the runaway sessions into a separate database. Redis Labs helped migrate data with a Lua script and Redis Cloud’s auto-scaling shards allowed capacity to grow (to 400GB on request), eliminating the leak, lowering backend load and restoring stable, high-performance real-time operations.

Glide

Max Rabin

Head of Server Development