Ready for your busiest day: How we scale

Fin, a platform serving some of the world's largest B2B and B2C companies, emphasizes its infrastructure's ability to handle significant scaling challenges, particularly during high-traffic events like Black Friday and sporting events. The platform processes over 150,000 customer requests per second at peak times and has developed strategies to manage both aggregate platform traffic and individual customer spikes. Fin's architecture relies on established technologies, like AWS, to focus on customer-specific needs, enhancing expertise and scalability. The company has also adopted Vitess, managed by PlanetScale, for its database, enabling improved availability and reduced complexity without customer downtime. Search optimization and customer workload isolation are key areas where Fin has implemented strategies to ensure efficiency and fairness, using tools like AWS SQS fair queues and application-level guardrails. The introduction of the AI Agent Fin presents new scaling challenges, which are managed through cross-vendor failover and capacity isolation. Real production traffic offers insights beyond synthetic tests, aiding continuous improvement and adaptation. Fin's operational model prioritizes customer outcomes, rapid deployment, and minimizing scheduled maintenance, underscoring a commitment to reliability and customer satisfaction during critical moments.