This opt-in feature allows hub admins to cap the amount of compute resource a user can access over a given time period.
Compute quotas¶
Suppose you are running a workshop and would like to cap usage to a 4GiB RAM server for each user per day.
With compute quotas you can apply a usage cap of 4GiB x 24 hours = 96 GiB-hours over a rolling 1 day window for every member of a particular JupyterHub group.
Alternatively, you can also apply a blanket compute usage policy to all users irrespective of JupyterHub group memberships.
For more details on compute policy configuration, see the jupyterhub-usage-quotas documentation.
Usage dashboards¶
To accompany this feature, there is a user-facing dashboard available to end-users to keep track of usage. The home storage component is enabled by default, which shows the “used versus available” disk usage in a user’s home directory. The compute component is enabled if a compute quota policy is applied to the hub.
Grafana dashboards¶
If there is a compute policy applied to your hub, then you will find a Usage Quotas dashboard in the 2i2c Dashboards folder of your Grafana instance that will help you monitor user’s compute usage. For more information on Grafana, see Usage and cost monitoring