// the find
hashicorp/nomad
Nomad is an easy-to-use, flexible, and performant workload orchestrator that can deploy a mix of microservice, batch, containerized, and non-containerized applications. Nomad is easy to operate and scale and has native Consul and Vault integrations.
Nomad is HashiCorp's workload scheduler — think Kubernetes but without the opinion that everything must be a container. It handles Docker, raw binaries, JVM apps, and VMs under one scheduler, and runs as a single binary with no external coordination service required. It's aimed at teams who need to orchestrate heterogeneous workloads across on-prem and cloud without rebuilding their entire stack around containers first.
Single binary with embedded Raft for consensus — no etcd dependency, no separate coordination plane to babysit. The task driver model is genuinely useful: you can schedule a legacy Java app or a raw executable alongside Docker containers in the same job file, which Kubernetes cannot do without significant contortion. Proven at scale (10K+ node clusters in production), and the federation model for multi-region deployments is first-class, not bolted on. Native Vault integration for secrets injection means you're not gluing in an external sidecar pattern.
The BUSL-1.1 license is a real constraint — if you're building a managed service or embedding Nomad in a product, you need to negotiate with HashiCorp. The ecosystem is thin compared to Kubernetes: fewer operators, less community tooling, and if you need a specific integration someone has already built a Helm chart for in the K8s world, you're writing it yourself here. Consul dependency is optional in theory but practically required for production service discovery, which means you're really adopting two systems. The HCL job spec is expressive but diverges enough from every other config format that onboarding engineers who know YAML-heavy stacks takes real time.