Capsule Settings

To open the Capsule Settings, double click on a capsule or hover over a capsule and click the gear icon.

If there is an App Panel for the Pipeline, the Arguments section will not be editable.

At the top of the Capsule Settings view the capsule’s description by hovering over the information icon. Open the capsule by clicking Go to this capsule.

Compute Resources

Once a capsule is added to a pipeline, it will inherit the resources selected in the capsule's environment editor. The capsule settings menu can then be used to modify the resources allocated to the capsule, independent of those set in the environment editor.

Each CPU based capsule can be allocated a maximum of 192 cores and 4096GB RAM. GPU based capsules can be allocated a maximum of 8 GPUs, 192 cores and 4096GB RAM.

By default all capsules will run on dedicated machines. Pipelines can be configured to run on spot instances in the Pipeline Settings.

Arguments

Supplying run script arguments in a pipeline capsule will replace the arguments that are supplied by that capsule’s App Panel. It is recommended to create an App Panel for the Pipeline rather than supplying arguments in this way.

Arguments must be separated by spaces. If an argument includes a space, such as the title of a plot, it should be in quotes.

Additional Note: Git Installation

For a capsule to run successfully in a pipeline, git must be installed. This is dependent on the starter image that is used for each pipeline. If not installed, include it as a package in the capsule's environment editor.