Case study


Twitch‘s next generation social video and community platform is architected for incredible scale with over 100M+ users per month, and 2M+ concurrent viewers on the site. Redis extensively powers the chat application, handling several hundreds of thousands of users in a single chatroom with extremely low latencies. Redis Enterprise delivers high performance, highly available Redis with exceptional operational simplicity that helps Twitch focus on presenting the best possible experience to their users.

Twitch Defines the Next Generation High Scale User Experience

Twitch brings over 100 million community members together with up to 2+ million concurrent visitors watching and talking about video games from 1.7 million + broadcasters. The top-notch engineering team that handles Twitch’s web application has architected their website for extremely fast response times as well as high availability and reliability.

Twitch uses several open source technologies to handle their data back end including Redis, Cassandra, Elasticsearch, DynamoDB and others. Redis is critical to their application as a first responder database, handling their website-wide chat functionality. Chat rooms often scale up to 400,000+ users in a single chat room, and low latency, highly available chat is key to their user’s social experience.


  • Need extremely high throughput, low latency persistent data store for high volume website-wide chat application
  • Required operational simplicity, high availability and high reliability

Redis Enterprise Benefits

  • Zero operational hassle, no maintenance worries
  • True high availability – no outages, no latency issues
  • Low overhead enables easy extension of Redis uses


“Redis Labs eliminated the need for Twitch to build operational expertise around managing the nuances of running Redis in production.”

Looking to the Future with Redis Enterprise

As the infrastructure grows and scales, the Twitch engineering team plans to extend its use of Redis Enterprise even more. Redis’s simplicity and versatility make it easy to transition use cases from other data stores. As an example, the team was initially using Cassandra as a key-value store but found it to be too complex operationally, imposing too much overhead. With Redis Labs making much of Redis operations a breeze, they are transitioning additional use cases to Redis.

Download the Case Study