Workspace Expiry & Costs

Workspace Expiry

What happens when my workspace expires? The workspace is automatically terminated and all data on the root volume is permanently deleted. Study data (S3) is not affected.

Can I extend my workspace? Contact your administrator before the expiry date to request an extension.

What about my data? Back up anything important before expiry. Transfer files to your local machine, a study, or an S3 bucket.

Costs

Why am I being charged when my workspace is stopped? Storage costs continue even when stopped. Only compute costs stop. Terminate workspaces you no longer need to stop all charges.

Why don’t my costs match my calculations? Common reasons: 24-hour delay in cost reporting, storage costs not included in your estimate, data transfer costs, or workspace ran longer than expected.

How can I reduce GPU costs? Use GPU instances only for GPU-specific tasks. Use CPU instances for data preparation. Stop GPU instances immediately when training completes.

What happens when I run out of credits? You’ll receive notifications as you approach your limit. Contact your administrator to request additional credits. Workspaces may be stopped if credits are exhausted.

Making the Most of Compute Service

What’s the single best way to save money? Stop your workspace when not in use. A medium instance running 8 hrs/day costs ~$24/month vs ~$108/month running 24/7 — a 78% saving.

How should I choose my instance size? Start small. Light development → Small; Data analysis → Medium; Heavy computation → Large; ML training → GPU (only when actively training).

Where should I store important data? Use studies rather than your workspace root volume. Study data survives workspace termination and can be shared across workspaces.

For detailed pricing tables, see the Cost Estimates Reference.