Case Study: Coinbase achieves 1.2M requests-per-minute scale and 10x platform resilience with MongoDB

Scaling for the future of finance with MongoDB

Coinbase, founded in 2012, grew from a Bitcoin wallet into a global crypto platform serving 20+ million accounts and facilitating over $220 billion in trades. During the 2017 crypto boom the team saw backend traffic jump from a steady 15,000 requests per minute to peaks that ultimately reached 1.2 million RPM, while also facing constant security threats — forcing engineering to prioritize both reliability and protection.

By standardizing on MongoDB and AWS and iterating quickly — upgrading to WiredTiger, using multi-document ACID support, moving analytics to secondaries, adding advanced monitoring, deploying a Memcached query cache, sharding clusters with data-driven shard keys, and rapidly integrating identity controls — Coinbase scaled without outages. The changes cut scaling time from hours to minutes, boosted platform resilience roughly 10x, and enabled the company to absorb 1.2M RPM traffic while continuing rapid product development.

Coinbase

Luke Demi

Software Engineer