hashicorp/nomad ? reverse-engineered prompt

Build me a simple but powerful workload orchestrator like Nomad. I want a single app with a CLI, API, and a basic web UI that can schedule and run different kinds of jobs across a cluster, not just containers but also regular executables, batch jobs, Java apps, and virtual machines where that makes sense. It should feel easy to operate, work on Linux, macOS, and Windows, and be able to recover from node or app failures with leader election and replicated state so a small team could actually trust it.

Please make it support service style workloads and one off jobs, scaling, health checks, rolling updates, and placement based on available CPU, memory, and special hardware like GPUs. It should also support running across multiple regions or clouds, and have hooks for secrets and service discovery with Vault and Consul, plus an HTTP API people can automate against. If there are docs or examples needed, add those too, and look up current docs online if you need to.