Company
Date Published
Author
Marko Rajcevic
Word count
1368
Language
English
Hacker News points
None

Summary

The Fortune 500 retailer weathered a regional cloud outage in Texas that knocked out power for over 4.5 million homes and businesses, with its public cloud data center affected for four days, due to the geo-distribution of its YugabyteDB cluster, which maintained application availability during the massive regional failure. The retailer used flexible geo-distributed topologies to set up a multi-region architecture that met their specific use cases and deployment needs, including low latency batch inserts, handling massive concurrency and parallelism, and scaling linearly. Despite losing all nodes in Texas without warning, the YugabyteDB cluster stayed up thanks to its strong consistency, rebalancing, and auto-sharding features, which enabled rapid adaptation to the crisis. The database's ability to sharded data, maintain data consistency, and perform efficient failover allowed it to handle the outage with minimal human intervention, resulting in considerable DevOps gains for the customer and avoiding financial or reputational damage due to downtime.