Volcano is a Cloud Native Computing Foundation incubating project that provides a batch scheduling system for compute-intensive workloads running on Kubernetes. It brings unified scheduling to Kubernetes-native tasks and widely used computing frameworks, including Spark, Flink, TensorFlow, PyTorch, Ray, Argo Workflows, Kubeflow, Open MPI, Horovod, MindSpore, MXNet, and PaddlePaddle. Designed for AI, big data, and scientific computing, Volcano helps teams manage demanding jobs and make more effective use of cluster resources.

Its capabilities include multi-level queue management with resource quotas and task priorities, support for heterogeneous hardware such as GPUs and NPUs, and network-topology-aware scheduling for distributed training. Volcano also supports multi-cluster scheduling, online and offline workload colocation, and load-aware descheduling. Users can choose from scheduling policies such as Gang scheduling, Fair-Share, Binpack, DeviceShare, NUMA-aware scheduling, and Task Topology. The project provides documentation, community resources, and contribution guidance. Its ongoing development adds scheduling enhancements and support for evolving Kubernetes and AI infrastructure needs.