Documentation | Deploy

Deploy

DeployResource

Methods

client.deploy.create() ->

AgentexCloudDeploy

post/v5/agentex/deployments

Create a new deployment.

Submits a deployment request. The deployment will:

Parse and merge configuration files
Create a Kubernetes Job running helm install
Wait for deployment completion
Return results

Parameters

environment_config: str

manifest_file: str

build_id: Optional[str]

The build_id of the cloud build. Required if image_name and image_tag are not provided.

expires_at: Optional[Union[str, datetime]]

(format: date-time)

ISO 8601 expiry timestamp. Only valid for preview deployments. If omitted on a preview deployment, defaults to 8 hours from now. Previews are always ephemeral and always have an expires_at.

image_name: Optional[str]

Name of the image to deploy. Required if build_id is not provided.

image_tag: Optional[str]

Tag of the image to deploy. Required if build_id is not provided.

preview: Optional[

bool

]

When True, creates a preview deployment with a unique deployment-id suffix appended to the helm release name.

preview_label: Optional[str]

Non-unique grouping label for the preview (e.g. branch name, PR number). Persisted on the deployment record so callers can list all deploys for a given label via GET /v5/agentex/deployments?preview_label=X&limit=1 (get the latest). Sanitized to lowercase alphanumeric + hyphens for K8s DNS-label compatibility (max 30 characters after sanitization). Each deploy still gets a unique helm release name regardless of label, so concurrent redeploys never share K8s resources. Only valid when preview=True.

Returns

AgentexCloudDeploy

id: str

account_id: str

created_at:

datetime

(format: date-time)

created_by:

Identity

The identity that created the entity.

environment_config: str

manifest_file: str

namespace: str

status: str

build_id: Optional[str]

deploy_events: Optional[List[

AgentexCloudDeployEvent

]]

Kubernetes events for this deployment.

expires_at: Optional[datetime]

(format: date-time)

helm_release_name: Optional[str]

object: Optional[Literal["agentex_cloud_deploy"]]

(default: "agentex_cloud_deploy")

preview_label: Optional[str]

Request example

import os
from scale_gp_beta import SGPClient

client = SGPClient(
    api_key=os.environ.get("SGP_API_KEY"),  # This is the default and can be omitted
)
agentex_cloud_deploy = client.deploy.create(
    environment_config="environment_config",
    manifest_file="manifest_file",
)
print(agentex_cloud_deploy.id)

200Example

{
  "id": "id",
  "account_id": "account_id",
  "created_at": "2019-12-27T18:11:19.117Z",
  "created_by": {
    "id": "id",
    "type": "user",
    "object": "identity"
  },
  "environment_config": "environment_config",
  "manifest_file": "manifest_file",
  "namespace": "namespace",
  "status": "status",
  "build_id": "build_id",
  "deploy_events": [
    {
      "message": "message",
      "reason": "reason",
      "timestamp": "2019-12-27T18:11:19.117Z",
      "type": "type"
    }
  ],
  "expires_at": "2019-12-27T18:11:19.117Z",
  "helm_release_name": "helm_release_name",
  "object": "agentex_cloud_deploy",
  "preview_label": "preview_label"
}

client.deploy.retrieve() ->

AgentexCloudDeploy

get/v5/agentex/deployments/{deployment_id}

Get a deployment by its ID.

client.deploy.list() -> SyncCursorPage[

AgentexCloudDeploy

]

get/v5/agentex/deployments

List all deployments with pagination and optional filters.

client.deploy.logs(, ) ->

DeployLogsResponse

get/v5/agentex/deployments/{deployment_id}/logs

Get structured log lines for a deployment with cursor-based pagination.

The CLI can poll this endpoint to stream logs incrementally:

First call: no cursor
Subsequent calls: cursor=next_cursor from the previous response
Stop polling when the deployment reaches a terminal status

client.deploy.delete() ->

AgentexCloudDeploy

delete/v5/agentex/deployments/{deployment_id}

Soft-delete a deployment. Marks it as deleted and triggers K8s resource cleanup.

Domain types

class AgentexCloudDeploy: ...

class AgentexCloudDeployEvent: ...

Slim event representation for the API response.

class DeployLogsResponse: ...

Response containing structured deployment log lines with cursor-based pagination.

The CLI can poll this endpoint to stream logs incrementally:

First call: no after_id
Subsequent calls: after_id=next_cursor from previous response
Stop polling when has_more is False and the deployment reaches a terminal status