Docker Offload usage

Docker Offload is billed for the time your cloud sandbox runs, as part of your Docker Agentic Platform subscription. For subscription and payment information, see Signup and billing.

What you're billed for

Billing starts when your sandbox finishes being created and stops when the sandbox starts pausing, whether you run docker offload stop or the sandbox pauses on its own after a period without activity.

You aren't billed while a sandbox is paused. A paused sandbox keeps its containers, images, and volumes, so you can leave a workload in place between sessions without paying for idle time.

By default, a sandbox pauses after 5 minutes without activity. You can change that period with --idle-timeout.

Warning

--keep-alive turns off idle detection, so the sandbox keeps running, and billing continues, until you run docker offload stop. It still pauses if your machine sleeps past the sandbox's time to live, and after 24 hours. For more information, see Control when your sandbox pauses.

Some workloads keep a sandbox awake even when you've stopped using it, which means it keeps billing. A container running on the host network does this, and so does a large number of running containers. Run docker offload status to see what's keeping a sandbox awake, and docker offload stop when you're finished.

Monitor your usage

To see your running sandbox and its hourly price, open the Docker Agentic Platform Console.

Find your sandbox in the Console

A sandbox that Docker Offload creates appears in the Console under your account with a name that starts with offload-. Use that prefix to tell it apart from sandboxes you started another way.

Your usage shows the sandbox's details, such as its size, vCPUs, and memory. It doesn't report that Docker Offload created the sandbox, so the name prefix is what identifies it.

Reduce your usage

Because you're billed for the time your sandbox runs, the main ways to reduce your spend are to stop your sandbox when you're not using it and to make your builds faster. For build and transfer optimizations, see Optimize Docker Offload usage.