Does my data leave my AWS account?
Saturn receives three pieces of information from users:
- Email Requests. Password reset emails are implemented as an HTTP request, which trigger our central service to send out an email.
- Email Addresses. We collect the email addresses of all Saturn users so we can notify you of upgrades, service disruptions, etc.
- General Usage Information. We collect pod usage records, which are required to calculate your monthly bill based on our pay-per-use pricing model. We pass usage records to the AWS Marketplace for the purposes of billing and disbursements.
NoteYour code and data stay inside your AWS account. That is never sent to us. This works because all of Saturn runs inside an EKS cluster, running inside your AWS account.
I just installed a package. Why can’t I import it?
A little bit about Jupyter
Environments in Saturn
In Saturn, we separate the environment for Jupyter, from the environment for the kernel. We do this so that you can do whatever you want in the kernel, without negatively impacting Jupyter. Jupyter is installed in /srv/conda, and your kernel, is installed in /srv/conda/envs/saturn . If you execute the folowing in a Jupyter notebook, you can verify which environment houses your kernel.d
>>> import sys >>> print(sys.prefix) '/srv/conda/envs/saturn`
if you’re installing packages with pip, the pip you’re using determines which environment is modified. If you execute /srv/conda/bin/pip you’ll end up modifying the environment that houses Jupyter. if you execute /srv/conda/envs/saturn/bin/pip you’ll end up modifying the environment that houses the Saturn kernel.
Conda is only installed in the base environment /srv/conda. With conda, you can select which environment to install into using the -n parameter (for name).
conda install -p /srv/conda will install packages into the base environment which houses Jupyter.
conda install -p /srv/conda/envs/saturn will install packages into the environment which houses the Saturn kernel.
It’s all in Docker
Because everything runs in Docker, if you modify the environment, your changes will not persist the next time you start Jupyter back up. The changes will also not be present on your Dask workers. As a result, we recommend building new docker images.
Where does Saturn store files with its version control feature?
Saturn provisions as Gitea deployment in EKS as part of its installation. Data is stored in an EBS volume in your AWS account
Why does Saturn Cloud run on Kubernetes?
Kubernetes is the most widely used container orchestration platform allowing to deploy, scale and manage containerized applications with a vast ecosystem of tooling to streamline infrastructure configuration and application development.
Why Amazon EKS?
EKS is a fully managed Kubernetes service offered by AWS. Saturn Cloud runs as an application inside Kubernetes leveraging AWS services such as EC2, AWS Identity and Access Management (IAM) and Amazon Virtual Private Cloud (VPC) to provide secure and scalable infrastructure to run your Data Science workloads.
Amazon EKS Pricing
Installing and using Saturn requires running a single Amazon EKS cluster that costs $0.10 per hour(~$72 per month). Multiple users within your organization can use the same Saturn installation running in a single EKS cluster .
You can find more Amazon EKS pricing details here https://aws.amazon.com/eks/pricing/
If you decide you no longer want to use Saturn, please send us an email at email@example.com, and we’ll tear down the deployment for you. Please do not delete the CloudFormation stack which provisioned the customer-admin IAM role, or delete the customer-admin IAM role, as we need that to tear down your resources.
The Saturn install process provisions many items in your AWS account. These include
- NAT Gateway
- Autoscaling Groups
- EKS Cluster
Do not delete the
customer-admin role we use to administer your installation. Shoot us an email, and we’ll take care of it and make sure there is nothing left behind.