An enterprise-grade, asynchronous job scheduler built without external dependencies. Native cgroups v2 isolation, DAG pipelines, and atomic NFS locks for heavy infrastructure.
$ nexus run --boost --timeout 3600 "./heavy_script.sh"🚀 Trabajo encolado con ID: A_1789067016-17495
Traditional schedulers (like Slurm or PBS Pro) consume excessive resources and require complex database setups. Nexus takes a minimalist approach by delegating state to a file-based spooling system.
It prevents out-of-memory (OOM) crashes by dynamically calculating the maximum allowed RAM, avoids network race conditions using atomic locks, and destroys hanging processes through strict timeouts.
Nexus is an open-source project maintained in my free time. If this tool has helped stabilize your servers, consider supporting its future.
1LUw27B6wPNibZKrPJGCigV642358rpABg
0x22a1dd7cd753ba5c52db23d927fc6ded7025579d