Teams

Coiled helps individuals and teams manage their resources, control costs, and collaborate with one another.

Using a Team account

If you are part of a team, you can create clusters or any other resources in your team account by using the account= keyword argument that most of our API commands accept, or by specifying my-team-account-name/ as a prefix in the name= keyword argument.

For example, to create a cluster in your team account:

import coiled

cluster = coiled.Cluster(n_workers=10, account="my-team-account-name")

Or, to create a software environment in your team account:

import coiled

coiled.create_software_environment(
    name="my-team-account-name/my-pip-env",
    pip=["dask[complete]", "xarray==0.15.1", "numba"],
)

Alternatively, you can also update your local coiled configuration file and add your team account name in this file to always use your team account when you are using coiled.

Sharing

Software environments and cluster configurations which belong to an account are visible and accessible to all account members. This allows team members to easily control, share, and collaborate on their teams’s resources.

Accounts & teams

Upon signing up, each Coiled user is automatically given their own personal account (the account name is the same as their username) which they can use to manage software environments, create cluster configurations, and launch Dask clusters.

Additionally, organizations and other teams can create their own Coiled account and add existing Coiled users to their team. Team members can be added and removed to an account at https://cloud.coiled.io/users. This enables team members to create software environments, cluster configurations, etc. within a team’s account (instead of their personal account) to easily share resources and collaborate with other team members.

../_images/team-management.png

Resource limits & tracking costs

Administrators for each Coiled account can set resource limits for account members like the number of cores a user can allocate at one time or whether or not to grant access to GPUs (which can be expensive). Additionally, you can track each cluster’s cost over time.

../_images/clusters-table.png