Company
Date Published
Author
Yiftach Shoolman
Word count
693
Language
English
Hacker News points
None

Summary

The Redis DevOps team experienced a significant outage in the AWS eu-central region due to an issue with one of the data centers. The outage affected multiple clusters, but thanks to their multi-AZ database configuration, all customers were unaffected and operations continued without significant downtime. This was made possible by several key principles, including in-memory replication, master and replica instances deployed across different zones, external persistent memory for fast recovery, uneven node deployment, and careful zone balancing. The Redis Enterprise Cloud architecture allows for an industry-leading five-nines availability, making it suitable for mission-critical use cases requiring sub-millisecond latency at high throughput.