Canonical MLOps
Canonical MLOps is a free, open-source platform for taking machine learning models from experimentation through production.
What Canonical MLOps is
Canonical MLOps is a free, open-source platform for taking machine learning models from experimentation through production. It combines Charmed Kubeflow, Charmed MLflow, and Charmed Feast across the machine-learning lifecycle. Kubeflow deploys, scales, and manages AI workflows on clouds, virtual machines, or bare metal. MLflow provides experiment tracking, reproducible projects, a model registry, and model deployment, and can run on a workstation, CNCF-conformant Kubernetes, or a public cloud. Feast supplies a centralized feature registry, offline and online stores, real-time and batch ingestion, and pipeline versioning. Automated workflows, security patching, and tooling integrations are included, with Jupyter, Grafana, and Prometheus listed for MLflow use. Kubeflow supports AWS, Azure, and GCP Kubernetes services. Canonical says components receive up to 15 years of security maintenance, updates, and CVE fixes.
Good
- Free plan available
- Includes Kubeflow, MLflow, and Feast
- Supports experiment tracking and model registry
- Works across clouds and bare metal
- Up to 15 years security maintenance
Watch out
- Designed around Linux, self-hosted, and web platforms
Who it suits
It suits teams running machine-learning workflows across Linux, self-hosted environments, or the web. It is also relevant for organisations needing experiment tracking, model registries, orchestration, serving, and cloud Kubernetes support.
Verdict
Canonical MLOps brings lifecycle tooling into one open-source stack, from features and experiments to deployment. Its broad deployment options and stated maintenance period may suit teams managing models across varied infrastructure.
Plans and prices
From Canonical MLOps's own pricing page, checked 5 Oct 2001.
| Plan | Price | What it gives you |
|---|---|---|
| Canonical MLOps support subscription | — | Open source stack · security maintenance · optional 24/7 support with SLAs · Per node, per year; price not stated on the page |
The facts
Each one read from Canonical MLOps's own pages, with the date it was checked.
What it does
- Automation
- The platform provides automated workflows, security patching, and tooling integrations for deploying and maintaining models.checked 5 Oct 2001
- Cloud support
- Charmed Kubeflow supports AWS, Azure, and GCP Kubernetes services, including EKS, AKS, and GKE.checked 5 Oct 2001
- Deployment
- Charmed MLflow can run on a workstation, CNCF-conformant Kubernetes, or a public cloud.checked 5 Oct 2001
- Feast features
- Charmed Feast provides a centralized feature registry, offline and online stores, real-time and batch ingestion, and versioning for ML pipelines.checked 5 Oct 2001
- Integrations
- Canonical lists Jupyter, Grafana, and Prometheus as tools to use with Charmed MLflow.checked 5 Oct 2001
- Kubeflow
- Charmed Kubeflow deploys, scales, and manages AI workflows across clouds, virtual machines, or bare metal.checked 5 Oct 2001
- MLflow features
- Charmed MLflow supports experiment tracking, reproducible projects, a model registry, and model deployment.checked 5 Oct 2001
- Pricing limit
- The Canonical MLOps page describes a per-node, per-year subscription but does not state a price.checked 5 Oct 2001
- Purpose
- Canonical MLOps is an open source platform for moving machine learning models from experimentation to production across the ML lifecycle.checked 5 Oct 2001
- Security
- Canonical says its MLOps components receive up to 15 years of security maintenance, regular updates, and CVE fixes.checked 5 Oct 2001
- Stack components
- Canonical’s MLOps stack includes Charmed Kubeflow, Charmed MLflow, and Charmed Feast.checked 5 Oct 2001
- Support
- Optional 24/7 support with SLAs is available for Canonical’s MLOps stack.checked 5 Oct 2001
The company
- Company
- Canonical says it has helped build the future of open source since 2004.checked 5 Oct 2001
- Founded
- 2004checked 28 Sep 2001
- Headquarters
- London, United Kingdomchecked 28 Sep 2001
Alternatives to Canonical MLOps
- 1
Databricks Notebooks
Free planAndroidWebAPILinux8.4Free plan - 2
Weights & Biases
Free planAndroidiOSWebAPI8.2From $60 a month - 3
Flyte
Free planOpen sourceWebAPILinuxSelf-hosted8.0From $950 a month - 4
Metaflow
Free planOpen sourceWebAPILinuxmacOS7.5Free plan - 5
ClearML
Free planWebAPILinuxmacOS7.3From $15 a month - 6
Opik
Free planWebAPILinuxSelf-hosted7.1From $19 a month