Serving status

Promotion status for a pipeline

The one call that answers "is this pipeline live, and what could I deploy instead?" — current deployment, promotion history, and every completed training in one response:

curl "$API_BASE/v1/deployments/pipelines/$PIPELINE_ID/status" \
  -H "Authorization: Bearer $TOKEN" \
  -H "X-Workspace-Id: $WORKSPACE_ID"

from predictai import PredictAI

client = PredictAI(token="pa_live_…", workspace_id="ws_…")

data = client.deployments.get_deployments_pipelines_by_pipeline_id_status(PIPELINE_ID)

import { PredictAI } from "@predictai/sdk";

const client = new PredictAI({ token: "pa_live_…", workspaceId: "ws_…" });

const data = await client.deployments.getDeploymentsPipelinesByPipelineIdStatus(PIPELINE_ID);

{
  "pipeline_id": "9f2c41d8-…",
  "is_promoted": true,
  "current_deployment": {
    "training_id": "b81e57c3-…",
    "deployment_type": "shared",
    "status": "deployed",
    "endpoint": "…/v1/models/model-b81e57c3",
    "created_at": "2026-07-15T12:04:31Z",
    "last_used": "2026-07-15T14:41:19Z"
  },
  "promotions": [
    {
      "promotion_id": "7d3f92a1-…",
      "training_id": "b81e57c3-…",
      "status": "active",
      "promoted_by": "usr_91c2…",
      "promoted_at": "2026-07-15T12:04:31Z"
    }
  ],
  "available_trainings": [
    {
      "training_id": "b81e57c3-…",
      "accuracy": 0.94,
      "created_at": "2026-07-15T09:12:44Z",
      "is_deployed": true
    },
    {
      "training_id": "a4c9d0e7-…",
      "accuracy": 0.91,
      "created_at": "2026-07-08T09:10:02Z",
      "is_deployed": false
    }
  ]
}

Response fields

Field	Meaning
`is_promoted`	`false` when nothing is serving — `current_deployment` is `null` then.
`current_deployment`	The most recent deployment. `status` here is a simple binary — `deployed` when ready, `deploying` otherwise. For the full lifecycle detail, use the deployment detail route.
`promotions`	The pipeline's deployment history; each entry shows as `active` or `deploying`.
`available_trainings`	Every completed training with its accuracy and whether it's currently serving — pick one and deploy it.

Everything currently serving

A flat list of live deployments, filterable by mode and status:

curl "$API_BASE/v1/deployments/serving?deployment_type=shared&status=deployed" \
  -H "Authorization: Bearer $TOKEN" \
  -H "X-Workspace-Id: $WORKSPACE_ID"

from predictai import PredictAI

client = PredictAI(token="pa_live_…", workspace_id="ws_…")

data = client.deployments.get_deployments_serving(params={"deployment_type": "shared", "status": "deployed"})

import { PredictAI } from "@predictai/sdk";

const client = new PredictAI({ token: "pa_live_…", workspaceId: "ws_…" });

const data = await client.deployments.getDeploymentsServing({
  params: { deployment_type: "shared", status: "deployed" },
});

{
  "deployments": [
    {
      "deployment_id": "7d3f92a1-…",
      "pipeline_id": "9f2c41d8-…",
      "model_name": "Daily sales forecaster",
      "workspace_id": "ws_3fa8…",
      "training_id": "b81e57c3-…",
      "deployment_type": "shared",
      "status": "deployed",
      "endpoint": "…/v1/models/model-b81e57c3",
      "resources": { "cpu": "200m", "memory": "512Mi" },
      "metrics": {
        "total_predictions": 0,
        "avg_latency_ms": 0,
        "last_24h_predictions": 0
      },
      "created_at": "2026-07-15T12:04:31Z",
      "last_used": "2026-07-15T14:41:19Z"
    }
  ],
  "total": 1,
  "page": 1,
  "pages": 1
}

Query parameters

Parameter	Meaning
`deployment_type` optional	`shared` or `dedicated`.
`status` optional	`deployed` or `deploying`.
`workspace_id` optional	Restrict to one workspace.
`page` optional	Page number. Defaults to `1`.
`limit` optional	Items per page.

Shared serving status

How the Shared pool looks right now — how many models are loaded and which ones:

curl "$API_BASE/v1/deployments/shared/status" \
  -H "Authorization: Bearer $TOKEN" \
  -H "X-Workspace-Id: $WORKSPACE_ID"

from predictai import PredictAI

client = PredictAI(token="pa_live_…", workspace_id="ws_…")

data = client.deployments.get_deployments_shared_status()

import { PredictAI } from "@predictai/sdk";

const client = new PredictAI({ token: "pa_live_…", workspaceId: "ws_…" });

const data = await client.deployments.getDeploymentsSharedStatus();

{
  "loaded_models": 12,
  "models": ["model-b81e57c3", "model-a4c9d0e7", …],
  …
}

Dedicated serving status

Every Dedicated deployment and its readiness:

curl "$API_BASE/v1/deployments/dedicated/status" \
  -H "Authorization: Bearer $TOKEN" \
  -H "X-Workspace-Id: $WORKSPACE_ID"

from predictai import PredictAI

client = PredictAI(token="pa_live_…", workspace_id="ws_…")

data = client.deployments.get_deployments_dedicated_status()

import { PredictAI } from "@predictai/sdk";

const client = new PredictAI({ token: "pa_live_…", workspaceId: "ws_…" });

const data = await client.deployments.getDeploymentsDedicatedStatus();

{
  "total_deployments": 2,
  "ready_deployments": 2,
  "deployments": [
    {
      "deployment_id": "3ac1f0b9-…",
      "ready": true,
      …
    }
  ]
}

If ready_deployments lags total_deployments, a deployment is still coming up or needs a look — inspect it and consider scaling or a retry.

Once a deployment reports ready, run real forecasts through Inference — segment-aware inputs, stable pipeline and goal URLs, and history.