Skip to main content

Managing Deployments

After you create a deployment, open it from the deployments list to inspect status, stream logs, open a shell, pause or resume it, clone it, or (for revision-aware types) roll back to an earlier spec.

This page covers those shared operations. Create-form fields stay on the type pages: LLM Serving, General Inference, NVIDIA Dynamo, Compute, and Batch Jobs.

Detail page tabs​

TabWhat it shows
DetailEndpoint, hardware, scaling, and (when supported) revision history.
LogsContainer stdout/stderr for a selected pod and revision.
TerminalAn interactive shell into a selected pod. Available while the deployment is Active.
UsageResource usage for the deployment.

Jobs have no serving endpoint. Use Detail and Logs (and Terminal while pods are running) to follow the batch run.

In-browser Terminal​

The Terminal tab opens a shell in the browser to a pod in an Active deployment. If more than one pod is running, choose the pod from the dropdown.

Use the Terminal for a quick look at processes, files, and GPU state without leaving the console.

The Terminal is available only while the deployment status is Active. For batch jobs, it is available while job pods are running. It disappears when the deployment is paused, completed, or no longer Active.

The shell you get depends on the image, the same as SSH on Compute:

ImageShell in the Terminal
Images with bash (for example Ubuntu, Debian, NGC images)bash
Images with only sh (for example Alpine)/bin/sh
Images without a shell (for example distroless)No interactive shell

For Compute instances, the Terminal does not replace SSH. Use SSH from your machine when you need scp/SFTP, port forwarding, or a persistent local session. Use the in-console Terminal for a short interactive session.

Pause and resume​

On the deployment detail page you can Pause an Active deployment or Resume a paused one. Pausing stops serving (or stops the Compute instance) without deleting the deployment.

  • LLM Serving, General Inference, Compute: pause and resume the whole deployment.
  • Dynamo: pause and resume the whole deployment. You cannot pause only an in-progress Dynamo rollout. See NVIDIA Dynamo.
  • Jobs: jobs run to completion and then stop. Pause and resume do not apply.

You can also pause and resume from the CLI (ccluster pause / ccluster resume). See Client Setup.

Clone and edit​

Clone opens the create form prefilled from an existing deployment (the name typically gets a -clone suffix). Review hardware, secrets, and tokens before you deploy the copy — cloned credentials such as Hugging Face tokens may need to be selected again.

Edit (update) changes the running deployment. On revision-aware types, spec edits create a new revision and start a rollout. Compute and jobs are not revision-aware in the console: an edit updates the deployment without a revision table or rollback UI.

Revisions, rollback, and rollout​

LLM Serving, General Inference, and Dynamo are revision-aware. The Detail tab lists revisions so you can inspect status and logs per revision and roll back to an earlier revision.

Compute and Jobs do not expose a revision table or rollback in the console.

When you edit a revision-aware deployment that is already Active, you can set rollout surge:

SettingDescription
Max surgeExtra replicas allowed above the desired count during the rollout.
Max unavailableReplicas that may be unavailable during the rollout.

Neither value can exceed the deployment's max replica count. Dynamo rollouts are operator-managed; see NVIDIA Dynamo for Dynamo-specific behavior.

Scaling replicas does not create a revision​

Changing Min Replicas or Max Replicas (scaling up or down) updates the running deployment in place and does not create a new revision.

A new revision is created when you change the deployment spec — for example image, command, environment variables, session affinity, model caching, or a config file — on a revision-aware type.

Rollback restores an earlier spec revision. It does not undo a replica-count change. To undo a scale-up or scale-down, set the replica fields back to the values you want.

Logs​

Logs appear on the Logs tab when Enable logging was on for that revision (or for the deployment, on types without revisions). Historical revisions keep the logs collected while they ran. If logging was off, the Logs tab has nothing to show for that revision.

On Dynamo, pick a Frontend or Worker pod; replica counts are reported per component. See NVIDIA Dynamo.

What each deployment type supports​

ActionLLM ServingGeneral InferenceDynamoComputeJobs
LogsYes, if logging is enabledYes, if logging is enabledYes, if logging is enabledYes, if logging is enabledYes, if logging is enabled
TerminalWhile ActiveWhile ActiveWhile ActiveWhile ActiveWhile pods are running
Pause / resumeYesYesWhole deployment onlyYesNot applicable
CloneYesYesYesYesYes
EditNew revisionNew revisionNew revisionIn-place updateIn-place update
Revision table / rollbackYesYesYesNoNo
Change rollout (max_surge / max_unavailable)YesYesOperator-managedNoNo
Scale replicas without a new revisionYesYesYesNot a multi-replica serving deploymentNot applicable

What's next​