Languages supported by Saturn Cloud
Run Python code with built-in support for installing packages with pip, conda and mamba.
Use R with packages installed from CRAN, GitHub etc.
Start an instance with full Julia support in seconds.
Any language that can be installed in Linux environment can be run in Saturn Cloud.
IDEs supported by Saturn Cloud
JupyterLab is built into Saturn Cloud for working with Python. JupyterLab enables you to work with Jupyter notebooks, terminals, and Dask clusters. You can also use JupyterLab for other languages.
Use RStudio Server to interact with Saturn Cloud when writing R code. RStudio Server lets you use R scripts, RMarkdown files, and more. You can also call Python from within R and RStudio with `reticulate`.
VSCode/Pycharm can connect to resources using the `Remote - SSH` plugin. VSCode will connect from your machine directly to the resource, bypassing JupyterLab or RStudio server.
Direct SSH terminal
Resources can be directly connected to from local terminals via SSH connections.
- Memory: 4GB RAM - 4TB RAM
- GPU: None, T4 GPU, or V100 GPUs (including multiple GPUs per machine)
- Hard Disk Space: 10GB - 1TB
- Logging: record your resource activity
- Utilization: dashboards for the resource and Dask
- Control: limit users and modify running resources
Connect your data stored in any AWS resource to Saturn Cloud via AWS credentials. You can also connect to other database platforms like MongoDB, Atlas, or Firebase.
Query your data in Snowflake and pull into Saturn Cloud through the Snowflake Connector for Python.
Access databases by setting up a start script for installing unixODBC for R or pyodbc for Python jobs.
Administative tools allow data science managers to Add Kaggle API key information to credentials page in Saturn Cloud and download Kaggle dataset to your current path.
Load local files by uploading JupyterLab or RStudio, or via SSH.
Saturn Cloud Backend
Users can use Saturn Hosted for Cloud Deployments. Also users can deploy Saturn in their own AWS environment with Saturn Cloud Enterprise , in which case we use the AWS Marketplace to install and manage the deployment.
Saturn Cloud runs as a deployment inside Amazon’s Elastic Kubernetes Service (EKS). EKS is a fully-managed Kubernetes service and is certified Kubernetes conformant. Each resource is deployed as a Docker container with the Kubernetes service.
Backup and Disaster Recovery
The Saturn Cloud application is automatically backed up nightly. This is preconfigured for all deployments and cannot be altered.
- Prefect Core, which is an open source library users can install.
- Prefect Cloud, which is a fully hosted cloud service which you can purchase.
- Unlimited resources - have as many resources as you want.
- Unlimited concurrent running resources - a user may have as many resources running at the same time as they want.
- Pay as you go - only pay for what you use.
- Isolation - each resource is totally independent of the others
- Scalability - resource size can be adjusted based on need
- Connected Clusters - Dask clusters can be connected for distributed computing
An image contains instructions that explain all the libraries and tools you want installed before you start work. Saturn Cloud has a standard CPU image as well as GPU instances with different sets of libraries like RAPIDS or PyTorch.
Minimal R and Python images to customize for your own use.
PyTorch Python GPU, TensorFlow Python GPU and RAPIDS Python GPU for machine learning.
Python Geospatial CPU, Python Snowflake CPU, and others for complex tasks.
Collaborating with colleagues
- Duplicate resources between colleagues so you can immediately run the same code.
Use groups to manage who has access to shared resources like production deployments.
Use recipes (a special kind of JSON file) to save resource specifications and keep code reproducible.