Open source Fly.io alternative

There are roughly 12-18 viable open source PaaS alternatives scattered across the landscape (Railway, Dokku, Coolify, Caprover, Render's open components, and various Kubernetes-based DIY stacks), but the market remains undersaturated relative to demand. The core pain points are fragmented: developers want Fly's DX without their pricing opacity, want to avoid AWS complexity, and want the ability to run on their own hardware or multi-cloud without rewriting their app. The gap isn't in the technology—several solid options exist—but in opinionated developer experience and a clear, honest positioning. Most open source PaaS projects struggle with documentation quality, unclear monetization paths (making founders nervous about adoption), and weak community momentum. The real opportunity sits at the intersection of three things: (1) making self-hosting stupidly easy with a single config file, (2) providing a transparent, developer-friendly pricing model that doesn't nickel-and-dime, and (3) building a thriving community around it. A new entrant should ruthlessly focus on one specific use case—either serverless functions, full-stack apps, or containerized microservices—and own it completely rather than trying to be Fly.io's perfect clone. The market reward goes to whoever makes the deployment + scaling + observability loop feel as native as pushing to GitHub.