Principal Engineer II at Razorpay - building Agent Studio, Ex-staff engg at GCP Memorystore & Dataproc, Creator of DiceDB, ex-Amazon Fast Data, ex-Director of Engg. SRE and Data Engineering at Unacademy. I spark engineering curiosity through my no-fluff engineering videos on YouTube and my courses
Stripe's DocDB: Scaling MongoDB to 5 Million Queries Per Second
Mar 13, 2026 • 1 min read
Introduction to DocDB: Stripe’s Scaled MongoDB Layer
Stripe, a leading financial technology company, leverages MongoDB as its primary data store. To meet its demanding scale and reliability requirements, Stripe developed an internal data layer called DocDB (Document Database) on top of the MongoDB Community Edition. This custom layer enables Stripe to handle an astounding 5 million queries per second, manage over 5,000 collections distributed across 2,000+ shards, and achieve five nines (99.999%) of uptime.
DocDB’s sophisticated design facilitates critical operations like no-downtime database migrations and efficient handling of hot shards, addressing many common challenges in large-scale database deployments.
Stripe’s Core Requirements for DocDB
Stripe designed DocDB with several key requirements in mind for its data layer:
- Availability, Durability, and Performance: Fundamental pillars for any mission-critical system, especially in financial services.
- Minimal Feature Exposure: To prevent