Case Study: GoDaddy improves user experience and anomaly detection with Elastic machine learning

Improving the GoDaddy User Experience with Elastic Machine Learning

GoDaddy, a leading web hosting and domain company serving 17 million customers with 75 million domains and 10 million hosted sites, faced a scale problem: hundreds of disparate Elasticsearch clusters and 200,000+ messages per second across DNS, logs and business events made it hard to get holistic visibility into patch compliance, infrastructure health and site performance — all critical to keeping users engaged.

GoDaddy centralized its Elastic deployment under a dedicated team (now managing 60+ clusters, 700+ Docker containers and 270 TB of indexed data) and sent NetFlow, sFlow, RUM and other telemetry into Elasticsearch. Using Kibana dashboards and Elastic machine learning jobs (RUM-focused percentiles and aggregations plus anomaly thresholds), they automated detection of performance issues, cut through noisy data, improved detection and response for site performance and patching, and gained actionable insights into customer experience and business KPIs.

GoDaddy

Felix Gorodishter

Principal Architect