Freedom251 › GPU Cluster Management Software

Koordinator

Free plan APILinuxSelf-hosted

What Koordinator is

Koordinator is a free, self-hosted scheduling system for managing microservices, AI and big-data workloads on Kubernetes. It works alongside Kubernetes’ scheduler, adding priority and quality-of-service controls that let different workloads and pods share a node. Application profiling supports resource overcommitment while preserving QoS guarantees, and fine-grained orchestration and isolation help manage latency-sensitive services alongside batch jobs.

Teams can configure scheduling policies and profiles for workloads such as web services, Spark, Presto, TensorFlow and PyTorch. QoSManager adjusts pod isolation settings based on profiling, interference detection and SLO configuration. Koordinator also provides monitoring, troubleshooting and operations tools. Koord-Queue supports TFJob, PyTorchJob, Spark, Argo Workflow, Ray and native Kubernetes Jobs, with compatibility for Kueue’s AdmissionCheck API. YARN Copilot allows Hadoop YARN jobs to share node-level batch resources with Kubernetes pods under unified priority and QoS policies.

The project runs control-plane and node components in a cluster without invasive changes to native Kubernetes components. Installation uses Helm 3.5 or newer and requires Kubernetes 1.18 or newer. The project is under the Apache License 2.0 and is a CNCF Sandbox project.

Who it suits

Koordinator suits teams that operate Kubernetes clusters with mixed service, AI, big-data or batch workloads and want configurable priority, resource sharing and isolation. It can also suit teams combining Kubernetes jobs with supported queue frameworks or Hadoop YARN.

What it does well

Free plan for self-hosted deployment.
Adds priority and QoS controls alongside kube-scheduler.
Supports profiling-based resource overcommitment and isolation.
Queue support includes several workload frameworks and Kueue.
Includes monitoring, troubleshooting and operations tools.
Apache License 2.0 project in the CNCF Sandbox.

What to know first

Requires Kubernetes 1.18 or newer.
Helm installation requires version 3.5 or newer.
NRI resource management requires containerd 1.7.0 or newer with NRI enabled.
Version 1.8 is only partially supported on Kubernetes 1.20 and 1.22.

Limits

Koordinator v1.8 has only partial support on Kubernetes 1.20 and 1.22, though core co-location, QoS and scheduling continue to function. NRI mode resource management requires containerd 1.7.0 or newer with NRI enabled.

Getting started

Visit the Koordinator website and installation documentation.
Use a Kubernetes cluster running version 1.18 or newer.
Install Koordinator with Helm 3.5 or newer.
For NRI mode resource management, use containerd 1.7.0 or newer with NRI enabled.

Verdict

Choose Koordinator if you need self-hosted Kubernetes scheduling with workload priorities, QoS controls and resource sharing across services and batch jobs. Teams on Kubernetes 1.20 or 1.22 should account for the partial support noted for version 1.8; the core co-location, QoS and scheduling functions continue to work.

Questions

Is Koordinator free?

Yes. Its pricing model is free, with a free plan.

Where does Koordinator run?

It is self-hosted in a Kubernetes cluster and uses control-plane and node components.

What does Koordinator require?

The installation documentation requires Kubernetes 1.18 or newer and Helm 3.5 or newer. Linux kernel 4.19 or higher is recommended for the best experience.

Which job systems does it support?

Koord-Queue supports TFJob, PyTorchJob, Spark, Argo Workflow, Ray and native Kubernetes Jobs. It is compatible with Kueue’s AdmissionCheck API.

Can Koordinator work with Hadoop YARN?

YARN Copilot supports running Hadoop YARN jobs with Kubernetes pods using shared node-level batch resources and unified priority and QoS policies.

What license does Koordinator use?

The project is licensed under the Apache License, Version 2.0.

Facts from Koordinator's own pages

CNCF status: Koordinator was accepted to the CNCF Sandbox on April 16, 2024. · cncf.io, 1 Oct 2026
Community support: Community communication is available through the Kubernetes Slack koordinator channel and DingTalk group 33383887. · github.com, 1 Oct 2026
Compatibility limit: Koordinator v1.8 is only partially supported on Kubernetes 1.20 and 1.22 because some components require newer Kubernetes APIs, while core co-location, QoS, and scheduling continue to function. · koordinator.sh, 1 Oct 2026
Founded: 2022 · koordinator.sh, 28 Sept 2026
Governance: Koordinator is a Cloud Native Computing Foundation sandbox project. · koordinator.sh, 1 Oct 2026
Installation: Koordinator can be installed with Helm 3.5 or newer, and the installation documentation requires Kubernetes 1.18 or newer. · koordinator.sh, 1 Oct 2026
Integration model: It runs with control-plane and node components in a Kubernetes cluster without invasive modifications to native Kubernetes components. · koordinator.sh, 1 Oct 2026
Job scheduling: Its job-scheduling mechanisms support workloads in areas including big data, AI, audio and video. · koordinator.sh, 1 Oct 2026
Kubernetes compatibility: Koordinator is compatible with Kubernetes QoS and its scheduler is designed to enhance kube-scheduler rather than replace it. · koordinator.sh, 1 Oct 2026
Kubernetes requirement: Koordinator requires Kubernetes version 1.18 or newer. · koordinator.sh, 1 Oct 2026
License: The Koordinator project is licensed under the Apache License, Version 2.0. · github.com, 1 Oct 2026
Linux requirement: The documentation recommends a Linux kernel version of 4.19 or higher for the best experience. · koordinator.sh, 1 Oct 2026
NRI support: NRI mode resource management requires containerd 1.7.0 or newer with NRI enabled and is enabled by default in Koordinator. · koordinator.sh, 1 Oct 2026
Operations tooling: Koordinator includes tools for monitoring, troubleshooting and operations. · koordinator.sh, 1 Oct 2026
Purpose: Koordinator is a QoS-based scheduling system for efficient orchestration of microservices, AI and big-data workloads on Kubernetes. · koordinator.sh, 1 Oct 2026
QoS and priority: It provides priority and QoS mechanisms to co-locate different workload types and run different pods on one node. · koordinator.sh, 1 Oct 2026
QoS management: QoSManager dynamically tunes pod resource-isolation parameters using profiling, interference detection, and SLO configuration. · koordinator.sh, 1 Oct 2026
Queue integrations: Koord-Queue supports TFJob, PyTorchJob, Spark, Argo Workflow, Ray and native Kubernetes Jobs, and is compatible with Kueue's AdmissionCheck API. · koordinator.sh, 1 Oct 2026
Resource efficiency: It combines elastic resource quotas, pod packing, overcommitment, resource sharing, and container isolation to improve utilization. · koordinator.sh, 1 Oct 2026
Resource isolation: It provides fine-grained resource orchestration and isolation for latency-sensitive workloads and batch jobs. · koordinator.sh, 1 Oct 2026
Resource overcommitment: It supports resource overcommitment using application profiling while maintaining QoS guarantees. · koordinator.sh, 1 Oct 2026
Runtime integration: NRI mode resource management requires containerd 1.7.0 or newer with NRI enabled and is enabled by default in the documented configuration. · koordinator.sh, 1 Oct 2026
Scheduling policies: It provides customizable scheduling policies and profiles for workloads including Web Service, Spark, Presto, TensorFlow, and PyTorch. · koordinator.sh, 1 Oct 2026
Security controls: Koordinator webhooks use least-privilege service accounts, TLS with automatically managed certificates, input validation, rate limiting, and audit logging. · koordinator.sh, 1 Oct 2026

Where Koordinator stands on Freedom251

GPU Cluster Management Software#11 of 21