Build and use a private image

A private image is a reusable disk image, not a live fork and not a backup. Build it from a computer you control, then create new computers from its img_… identifier.

Prepare a clean source

Install the required packages and application files. Remove credentials, histories and application-specific machine identities before capture. The platform installs only the new computer's create env/setup as CE-managed configuration; arbitrary files and services baked into the image are not scrubbed. Do not assume image creation removes secrets for you.

Stop the source and wait for the stopped state before POST /images. Keep the accepted image operation ID and poll it to success. Only then create a new computer with that image ID and fresh environment/setup. A request that names an unknown or unauthorized image returns 404, not a cross-account disclosure.

Limits and cost

Image creation is limited to two per day on Tier 1 and ten on Tier 2, as rolling counts. Consult GET /limits before creating another image; image metadata and retained storage are separate from running-computer limits. Account storage includes private images, with the shared 10 GB allowance and $0.05 per GB-month beyond it. Images survive ordinary deletion of the source computer.

Retention and cleanup

Low-balance retention applies even to an account with only private images. Quarantined images cannot create new computers; read retention_state and scheduled_deletion_at. Legal hold and unresolved payment/refund evidence can prevent destruction. Export important source material outside the server before deleting the final copy. See Snapshots and images and Low balance.

Dory is in development; these pages describe the planned service and may change. Questions: support@computeengines.com.