Marathon

Marathon is a production-grade container orchestration platform designed to manage long-running services on Apache Mesos.

Marathon acts as a cluster-wide init system for containerized workloads, ensuring high availability through active-passive leader election and automated self-healing. It supports both Docker and Mesos container runtimes, allowing operators to scale applications to thousands of instances using a robust REST API or a clean web UI. By leveraging Apache ZooKeeper for state management, Marathon maintains 100% uptime for critical services like MySQL or Postgres (even during hardware failures) by automatically rescheduling tasks onto healthy nodes. Its mature feature set includes rolling upgrades, health checks (HTTP and TCP), and sophisticated constraint engines for precise workload placement across racks and datacenters.

https://mesosphere.github.io/marathon/