GitLabが5,000万ユーザーに対してサブ秒の分析を提供する方法

GitLab transitioned to ClickHouse for its analytics infrastructure to address the need for sub-second insights at large scales, moving away from using Postgres for both analytics and transactional workloads. This shift was driven by ClickHouse's ability to handle extensive data volumes with remarkable speed, reducing query times for over a billion rows from 30-40 seconds to under a second. ClickHouse's integration supports features like Contribution Analytics and GitLab Duo, allowing real-time tracking of engineering outputs and AI adoption. GitLab's strategic pivot included making ClickHouse the default OLAP database across all deployments, while Postgres continues to handle OLTP tasks. The adoption of ClickHouse Cloud reduced operational burdens, allowing teams to focus on delivering high-performance, scalable insights without being bogged down by infrastructure complexities. This move aligns with GitLab's broader strategy to embed analytics deeply into its platform, enhancing the ability to offer real-time, scalable insights across all customer environments.