On our JupyterHub, you get a personal JupyterLab container. By default, your own user directory and any additional group folders are included. The performance of these containers is limited to ensure that all users can use the service properly.
The current resource limits are:
- Memory: 32 GB
- Processors: 4 cores
- Maximum number of processes (PIDs): 1000
- Maximum number of open files (ulimits -n): 1024
We follow a fair use policy for storage space, i.e. there are no fixed quotas per user. This allows you to temporarily store larger amounts of data, for example, for intermediate results.
The use of the graphics card is also based on the fair use policy. This resource is shared by all JupyterLab instances and cannot be reserved per user. We ask you, especially when working with machine learning models, to release the video random access memory (VRAM) promptly after finishing your tasks to enable other users to use it efficiently.
Please note that inactive instances are automatically terminated to conserve resources. Instances are terminated after no more than 12 hours of runtime, regardless of their activity.