Majordodo is a distributed resource manager developed by Diennea engineering for building complex, multi-tenant computing grids. It works like a distributed thread pool: users submit tasks, and a network of brokers schedules them across a pool of machines. The platform is designed to manage execution priorities, resource allocation, failures, and recovery across distributed environments.

Tasks are implemented in Java and can be submitted and monitored through a straightforward HTTP REST JSON API. Majordodo supports multi-tenancy by managing resources separately for each tenant, providing access isolation and fair scheduling. Execution priorities can be changed at runtime for individual tasks, including tasks already in the scheduler.

For resilience, Majordodo uses Apache ZooKeeper and Apache BookKeeper to support a replicated, shared-nothing broker architecture without a single point of failure. Workers can run on commodity hardware, virtual machines, or dynamically provisioned cloud resources. Built to handle tasks across thousands of hosts, the system supports adding or removing hosts and resources through dynamic reconfiguration and adaptive task scheduling.