On January 29, there was unprecedented demand for Kraken services that far exceeded our surge prediction modeling.
For context, $56 billion was traded on our platform in January of this year – more than all the volume transacted for the full year of 2019. This past weekend alone, we saw a 1,000%+ increase in signups.
Unfortunately, all of this activity caused intermittent downtime and cascading issues with both the front-end and back-end systems that facilitate client experiences across our Web UI, REST API, and at times, our mobile apps and Websockets.
The Kraken Engineering team rallied to coordinate a massive improvement to our services. This included expanding capacity in all possible places and rolling out performance enhancements that had been in beta.
While we don’t typically expose what happens behind the scenes at Kraken because of our focus on security, we feel it is important to provide transparency by explaining to our clients what happened and what we’re doing to prevent this in the future.
We put in place new hardware and expanded capacity, deployed improvements to our systems and databases and further optimized our website and the sign-up process for new clients.
For our more technical clients, here are the specifics:
- Increased front-end nodes by 30% and replaced older nodes with brand-new hardware that is 3x more performant
- Optimized Kraken.com to improve the sign-up process for new clients
- Deployed optimized components for databases, session management and API calls
- Added new points of telemetry to aid in earlier detections of future spikes
- Deployed new service upgrades that greatly reduced latency on some of our most heavily used services from several seconds to 25 milliseconds
- Massively increased capacity to the systems and services involved in processing and verifying documents essential for new client sign-ups
- Upgraded the performance of our back-end support systems when working with clients to review…