Scale to one: How Fluid solves cold starts

This post introduces 'scale to one' as a strategy to keep at least one function instance warm, preventing first-visitor cold starts. It also details how Fluid compute reuses existing instances for multiple requests, reducing the number of cold starts. Predictive scaling is also mentioned as a method to warm instances ahead of demand. The post quantifies the reduction in cold starts to less than 1% of requests.