MLflow Tracking Server Self-Hosted
Self-hosted experiment tracking and LLM/agent observability with customer-controlled metadata and artifacts.
3 sources checked by an editor on · Product website ↗
Read this page as JSON or Markdown.
Deployment terms: BYOC · control plane · data plane · private offer
· Open requirements shortlist · Estimate operating costs
Compare this listing with another →
What lands where
Supplied by
Vendor or project
Python MLflow server command, official Docker Compose bundle or Kubernetes Helm deployment.
Runs in
Documented customer environment
Customer-operated tracking/API server with configurable metadata database and artifact storage
- AWS: Customer cloud account
- Azure: Customer cloud account
- Google Cloud: Customer cloud account
- AWS
- Azure
- Google Cloud
Servers / containers
Operated by
Your team
Customer operates server access, metadata and artifact stores, retention, backups, scaling and version upgrades.
Deployment boundaries by provider
- AWS: Customer cloud account · Customer-operated
Resources in the customer's cloud account, subscription, project or tenancy. Managed cloud services still have shared responsibilities.
Self-hosted MLflow deployment on AWS ↗ · Source checked 2026-11-12
- Azure: Customer cloud account · Customer-operated
Resources in the customer's cloud account, subscription, project or tenancy. Managed cloud services still have shared responsibilities.
Self-hosted MLflow deployment on Azure ↗ · Source checked 2026-11-12
- Google Cloud: Customer cloud account · Customer-operated
Resources in the customer's cloud account, subscription, project or tenancy. Managed cloud services still have shared responsibilities.
Self-hosted MLflow deployment on Google Cloud ↗ · Source checked 2026-11-12
Application operations and account ownership are separate from infrastructure operations. Confirm external services, telemetry, support access and backup destinations for the chosen path.
| Install method | Python MLflow server command, official Docker Compose bundle or Kubernetes Helm deployment. |
|---|---|
| Infrastructure | Servers / containers The tracking server runs as a Python service or container with metadata/artifact stores. A local SQLite example is not a production architecture; separately plan persistent database, storage and application-runtime infrastructure. |
| Licensing model | Open Source / Open Core The upstream self-hosted tracking server is Apache-2.0 licensed. This classification does not extend to separate managed MLflow distributions, connected model providers or their commercial services. |
| License | Apache 2.0 for upstream MLflow; connected model providers and managed distributions have separate terms. |
Buyer diligence
Answers apply to the reviewed edition and deployment. Expand a row for qualifications and sources.
How to read these facts
Not researched means no field-specific review is recorded. Not confirmed means sources were checked but did not establish an answer. Availability, eligibility and edition-dependent findings retain the scope given in their evidence. A vendor’s hosted-service certification does not automatically cover your deployment. Evidence coverage is not a feature score.
5 of 12 fields researched (42%) · 4 documented · 1 unresolved. Evidence coverage is not a feature score.
- Single sign-on
AvailableEvidence and scope for Single sign-on
The official guide documents the community-maintained mlflow-oidc-auth plugin and an SSO reverse-proxy pattern. This provides UI/API OIDC integration, but plugin maintenance, MLflow-version compatibility and client token behavior require separate review; it is not a claim of vendor-operated identity support.
Officially documented community-plugin/proxy integration · Source ↗ · Reviewed 2026-11-20
Documented · Editor reviewed.
- Deployment artifact / IaC
DocumentedEvidence and scope for Deployment artifact / IaC
The self-hosting guide links an upstream Compose bundle with MLflow, PostgreSQL and object storage, plus an official Helm option. Configure production credentials and persistence rather than retaining development defaults.
Upstream self-hosted deployment; not a provider support or production-readiness certification · Source ↗ · Deployment artifact ↗ (master Compose bundle (moving)) · Reviewed 2026-11-14
Documented · Editor reviewed.
- Minimum footprint / sizing
Not confirmedEvidence and scope for Minimum footprint / sizing
The self-hosted guide documents the UI/API tracking server, metadata backend and artifact-store configuration, but not a universal production CPU/RAM minimum. Size runs/traces, request concurrency, database, artifact storage and redundancy; starting the default local SQLite server does not establish production or HA capacity.
Self-hosted tracking-server topology; numeric production sizing unresolved · Source ↗ · Reviewed 2026-11-20
Sources checked — unresolved · Editor reviewed.
- License pricing
DocumentedEvidence and scope for License pricing
The reviewed upstream core is Apache-2.0 licensed. Optional services, bundled dependencies, commercial distributions and support can have separate terms. This records the source licence, not an infrastructure or operating-cost estimate.
Upstream core licence; commercial features, support and hosting priced separately · Source ↗ · Reviewed 2026-11-20
Documented · Editor reviewed.
- Release snapshot / cadence
DocumentedEvidence and scope for Release snapshot / cadence
The publisher-designated GitHub release checked was v3.18.0, published 2026-11-04. This is a release snapshot, not a promised cadence, support designation or approval to substitute it into a provider package.
Upstream mlflow/mlflow release record; commercial editions and deployment packages may differ · Source ↗ · Reviewed 2026-11-14
Documented · Editor reviewed.
Show all fields (7 not researched)
- Marketplace delivery
- Not researched Supply a public source.
- Cloud commitment eligibility
- Not researched Supply a public source.
- SCIM provisioning
- Not researched Supply a public source.
- Compliance attestations
- Not researched Supply a public source.
- Security controls
- Not researched Supply a public source.
- Air-gapped operation
- Not researched Supply a public source.
- Operating cost estimate
- Not researched Supply a public source.
Scope & considerations
Traces can capture model inputs, outputs and intermediate metadata. Point clients at the customer-hosted server and review recording/redaction settings. Applications, evaluators and optional model integrations may still call external providers; hosting the tracking server does not host all training or inference runtimes. Production database, storage and security configuration differ from a local quickstart.
Category fit
- AI observability & LLMOps
- Experiment tracking and LLM/agent tracing server. This entry describes customer-hosted tracking/observability services, not every model runtime, managed MLflow product or inference backend.
- AI tools & infrastructure
- Self-hosted experiment tracking and LLM/agent observability with customer-controlled metadata and artifacts.
- Observability
- Application/model traces and experiment visibility for AI workloads. Infrastructure metrics and broader application monitoring require separately selected tooling.
Compare within AI observability & LLMOps
- Arize Phoenix Self-HostedAI observability and evaluation server for traces, datasets and application experiments.
- AWS
- Azure
- Google Cloud
- Railway
- Render
- Langfuse Self-HostedSelf-hosted LLM observability, evaluations and prompt-management platform.
- AWS
- Azure
- Google Cloud
- Railway
- Render
- OpenLITOpenTelemetry-based AI tracing and evaluation platform for model and agent operations.
- DigitalOcean
These products address related workflows; read their category-fit notes before comparing scope. See all of AI observability & LLMOps →
Installation outline
- Choose the documented deployment option and a dedicated cloud environment. Follow the launch methodology.
- Review the vendor's installation documentation, required permissions and outbound data flows.
- Deploy to a non-production environment; configure identity, networking and monitoring.
- Test backup recovery and upgrades before promoting to production.
Evidence trail
Every claim on this page traces to the sources below. They document the deployment option; they do not constitute a security certification.
- Deployment sourcesChecked
- Self-hosting server and deployment methods ↗mlflow.org
- LLM/agent tracing and captured data ↗mlflow.org
- Apache 2.0 license ↗github.com
- Cloud deployment evidenceChecked
The linked sources describe each deployment path, edition and support scope. Community templates and partner statements are identified in the notes.
- Self-hosted MLflow deployment on AWS ↗ · AWS
Vendor / upstream documentation · Reviewed 2026-11-12. Vendor support commitment: not assessed; confirm the selected release and contract.
The project documents deployment of the MLflow server, metadata database and artifact storage on AWS. This does not refer to a managed MLflow edition or to model serving alone.
- Self-hosted MLflow deployment on Azure ↗ · Azure
Vendor / upstream documentation · Reviewed 2026-11-12. Vendor support commitment: not assessed; confirm the selected release and contract.
The project documents deployment of the MLflow server, metadata database and artifact storage on Azure. This does not refer to a managed MLflow edition or to model serving alone.
- Self-hosted MLflow deployment on Google Cloud ↗ · Google Cloud
Vendor / upstream documentation · Reviewed 2026-11-12. Vendor support commitment: not assessed; confirm the selected release and contract.
The project documents deployment of the MLflow server, metadata database and artifact storage on Google Cloud. This does not refer to a managed MLflow edition or to model serving alone.
- Self-hosted MLflow deployment on AWS ↗ · AWS
- Infrastructure evidenceChecked
- Licensing evidenceChecked