Deployments

Get the full configuration of a specific revision

Changed on

Returns the deployment configuration defined by the specified revision. Only the 50 most recent revisions per deployment retain their configuration; older revisions return 404 even while they still appear in the revision history. The deployment's currently active revision is always available, regardless of age.

get/v1/deployments/{id}/revisions/{revisionIdentifier}

Request

  • Base URL
    https://api.together.ai
  • URL
    https://api.together.ai/v1/deployments/{id}/revisions/{revisionIdentifier}
  • Auth
    HTTP bearer

Path parameters

idstring required

Deployment ID or name.

revisionIdentifierstring required

Revision number or revision ID.

Response

Configuration of the specified revision

object'revision' required

The object type, which is always revision.

revision_idstring required

Unique revision identifier.

imagestring required

Container image used by this revision.

min_replicasinteger required

Minimum number of replicas configured for this revision.

max_replicasinteger required

Maximum number of replicas configured for this revision.

portinteger required

Container port exposed by this revision.

protocolstring required

Network protocol served by the deployment revision.

capacity_type'stable' | 'preemptible' required

Capacity behavior for replicas above reserved capacity.

commandstring[] required

Entrypoint command run by the container.

argsstring[] required

Arguments passed to the container command.

gpu_typestring required

GPU hardware type configured for this revision.

gpu_countinteger required

Number of GPUs allocated to each replica.

cpunumber required

CPU cores allocated to each replica.

memorynumber required

Memory allocated to each replica in GiB.

storageinteger required

Ephemeral storage allocated to each replica.

health_check_pathstring required

HTTP path used for health checks.

termination_grace_period_secondsinteger

Seconds to wait for graceful shutdown before forcefully terminating a replica.

created_atstring date-time required

Time when this revision was created.

Example response

{
  "autoscaling": {
    "metric": "HTTPTotalRequests",
    "target": 100,
    "time_interval_minutes": 10
  }
}

Changes