Case Study: Facebook achieves scalable, resilient, developer-friendly search with Elastic (Elasticsearch)

Facebook - Customer Case Study

Facebook faced the need to replace brittle, hard-to-manage search tooling (Google Search Appliance and ad-hoc Solr setups) as an internal hackathon project grew into production. The challenge was to onboard engineers with little search experience, support dozens of product teams, and scale from a single small cluster to multiple clusters across datacenters while handling billions of documents, high query volume, and reliable migrations and disaster recovery.

They standardized on Elasticsearch for its REST/JSON API, Lucene power, and ecosystem, and built deployment and developer tooling around it: sandbox docs, index/mapping templates, containerized nodes (Tupperware/LXC), monitoring (Scuba), cross-cluster replication and alias-based cutovers, automated snapshots, and Shield for security. The result was rapid adoption, easier onboarding, and a production platform running 100+ nodes in multiple datacenters serving ~4 billion documents and 1,500+ QPS with repeatable migration and recovery processes.

Facebook

Peter Vulgaris

Facebook