Get a compute resource

Retrieve the configuration, capacity, and current status of a compute resource.

HTTP

GET $CLOUDSIGMA_API_BASE/compute-resources/{resource_id}

Before you call this API

1. Configure the regional API endpoint and authentication, and select the target workspace.

2. Obtain the compute resource ID from List compute resources or from the response returned when you create a compute resource.

3. Store the compute resource ID in an environment variable:

Shell

export RESOURCE_ID="<compute-resource-id>"

Request

Input examples

Shell

curl "$CLOUDSIGMA_API_BASE/compute-resources/$RESOURCE_ID" \
  -H "X-API-Key: $CLOUDSIGMA_API_KEY" \
  -H "X-Workspace-ID: $WORKSPACE_ID"

Parameter reference

Path parameters

Parameter	Type	Required	Description
resource_id	string	Yes	ID of the compute resource to retrieve.

Successful response

The API returns HTTP 200 and the compute resource in data.

Response example

JSON

{
  "code": "OK",
  "msg": "OK",
  "data": {
    "id": "cr-001",
    "name": "data-processing",
    "kind": "task",
    "spec_id": "task-standard-small",
    "status": "ACTIVE",
    "status_message": "",
    "min_replicas": 0,
    "max_replicas": 2,
    "current_replicas": 1
  }
}

Show full code

Field reference

Response envelope

Field	Type	Description
code	string	OK on success.
msg	string	Human-readable result message.
data	object	Retrieved compute resource.

Identity and configuration

Field	Type	Description
data.id	string	Compute resource ID.
data.workspace_id	string	ID of the workspace that owns the resource.
data.name	string	Compute resource name.
data.description	string	Compute resource description, when provided.
data.spec_id	string	Compute specification ID.
data.kind	string	Compute resource type.
data.cpu	integer	Allocated CPU capacity.
data.memory_gib	integer	Allocated memory in GiB.
data.gpu	integer	Allocated GPU capacity.
data.cpu_milli	integer	Allocated CPU capacity in millicores.
data.memory_mib	integer	Allocated memory in MiB.
data.gpu_count	integer	Number of allocated GPUs.
data.gpu_memory_mib	integer	Memory per GPU in MiB.
data.gpu_cores	integer	Number of GPU cores.
data.go_worker_image_id	string	Go worker image ID, when configured.
data.python_worker_image_id	string	Python worker image ID, when configured.
data.worker_images	object[]	Worker images assigned to the resource.
data.worker_images[].worker_type	string	Worker type.
data.worker_images[].image_id	string	Image ID.
data.worker_images[].platform	string	Worker image platform.
data.platform	string	Runtime platform, when provided.

Capacity and status

Field	Type	Description
data.min_replicas	integer	Minimum number of replicas. A serverless resource uses 0.
data.max_replicas	integer	Maximum number of replicas.
data.desired_replicas	integer	Requested number of replicas.
data.current_replicas	integer	Current number of replicas.
data.auto_suspend_minutes	integer	Inactivity period, in minutes, before automatic suspension.
data.is_default	boolean	Whether this is the workspace’s default compute resource.
data.status	string	Current lifecycle state, such as IDLE, ACTIVE, SUSPENDED, or ERROR.
data.status_message	string	Additional status information.
data.scale_reason	string	Reason for the most recent scaling decision, when provided.
data.last_activation_at	string	Time of the most recent activation, when available.
data.last_active_at	string	Time the resource was most recently active, when available.
data.created_by	string	Creator ID; returned if there is a value.
data.created_at	string	Creation time.
data.updated_at	string	Last update time.

In a field path, [] represents each item in an array. For example, data.worker_images[].worker_type is the worker_type of each object in data.worker_images.

Error response

Response example

JSON

{
  "code": "NOT_FOUND",
  "msg": "对象不存在",
  "data": null
}

Field reference

Common HTTP errors

HTTP status	Error code	Common cause	Recommended action
401	—	Missing or invalid access credentials.	Check the regional endpoint and API key.
403	ErrForbidden	The current identity does not have permission to read the compute resource.	Ask a workspace administrator to grant the required permission.
404	NOT_FOUND	The compute resource does not exist in the selected workspace.	Check RESOURCE_ID, WORKSPACE_ID, and the regional endpoint.

Follow-up operations

After retrieving the resource, you can:

• Get its worker runtime.

• Update its configuration.

• Suspend, resume, or retry the resource according to its current status.

Last updated on Sep 15, 2026