Responding to Growth - Increased Reliability and Reduction of Infrasture Costs
Summary
Unfuddle faced challenges with rapid growth and high traffic. Our team redesigned their infrastructure and rewrote massive portions of their web application. This resulted in 80% cost savings and improved service reliability.
Introduction
Since 2006, Unfuddle has provided Git and SVN hosting integrated with bug and issue tracking. Unfuddle were pioneers in their field, predating major platforms like GitHub. They serve thousands of companies globally, including many high-profile clients.
Challenge
Through a period of high growth, Unfuddle's infrastructure costs soared, reaching unsustainable amounts. Ensuring consistent uptime became increasingly difficult due to the high volume of requests. Monitoring the expanding infrastructure and building backup redundancy further compounded these issues. As is often the case, growth was becoming a liability, posing a significant risk to customer satisfaction and hindering potential business growth.
Solution
- Separated the web service into elastic clusters to enhance scalability and reliability.
- Migrated to the new infrastructure with minimal to no downtime, ensuring business continuity.
- Adapting to the unique demands of Unfuddle's offering, we developed proprietary tools for managing the old and new infrastructure, showcasing significant creativity and innovation.
- Reduced CPU usage on web servers by moving the user interface generation from the servers to the browser.
- Achieved a fully cached, performant browser application that was live and reactive, further reducing the need for web server resources.
Impact
- Resource costs reductions by a shocking 80%.
- Enhanced service reliability and uptime.
- Improved response times.
- Achieved significant reduction in CPU usage on web servers.
