Datalore is available as a privately hosted service for companies on the Enterprise plan.
Install Datalore on the Kubernetes cluster of your choice, be it on-premises or a private cloud. Kubernetes setup is the most customizable method, but it requires knowledge of Kubernetes technology. Find more information in our installation guide.
Here is how the Kubernetes setup looks:
With Docker-based installation you can get a basic, private Datalore setup running in less than 10 minutes. Whether it’s on AWS, GCP, Azure or an on-premises machine, you’ll be able to set up Datalore by running a single Docker command. Connecting authentication modules, setting up internal usage plans, and customizing environments can be further carried out step-by-step.
Use Docker Swarm to connect multiple machines to your Datalore installation. Docker Swarm machine connections can be used as an alternative to Datalore’s Kubernetes-based installation if you don’t have an existing k8s cluster.
Connect any machines to Datalore Kubernetes or Docker-based installations with TeamCity. This way, you can add machines outside your cluster to your Datalore computational power.
Datalore is available for hosting in networks that have no access to the internet.
Configure almost any possible method of user authentication to let your team seamlessly access Datalore. The Enterprise version includes JetBrains Hub – the module for connecting authentication providers.
Easily manage who can access notebooks and workspaces via group sharing. Create user groups inside Datalore’s Admin panel or sync them with the JetBrains Hub authentication module. JetBrains Hub can sync user groups for selected authentication modules with your authentication provider (e.g. Azure AD). Find more information on supported Auth modules here.
Create multiple base environments out of custom Docker images. You can pre-configure all the dependencies, package versions, and build tool configurations so that your team doesn’t spend time on installing things manually and syncing package versions.
Different members of your data science team might need different computational resources. Datalore allows you to create separate usage plans for those who require high-efficiency GPUs and for those who need a lot of regular CPU computations, for example. You can specify how many hours each user is allowed to run each machine type, and restrict some users from running high-cost machines.
You can turn on user action logging inside your on-premises setups. The logs will help you investigate any compliance requests.