Updated on 2026-09-03 GMT+08:00

Scaling In a Cluster by Removing a Specific Node

Function

If a cluster has excess capacity due to off-peak traffic or reduced data volume, you can reduce its nodes to cut costs.

Constraints

Data migration is involved during scale-in. The timeout threshold for data migration of a single node is 48 hours. If the timeout occurs, the scale-in fails. When the cluster data volume is large, manually adjust the data migration rate and avoid operations during peak hours.
Before scaling in, back up all key data to prevent data loss.
Only Elasticsearch/OpenSearch clusters are supported.
Data migration is involved during scale-in. The timeout threshold for data migration of a single node is 48 hours. If the timeout occurs, the scale-in fails. When the cluster data volume is large, manually adjust the data migration rate and avoid operations during peak hours.
For clusters without master nodes: scale-in is supported only when the total number of data nodes and cold data nodes is greater than or equal to 3. In a single scale-in operation, the total number of data nodes and cold data nodes to be removed must be less than half of the total number before scale-in. After scale-in, the total number of data nodes and cold data nodes must be greater than the maximum number of index replicas.
For clusters with master nodes: scale-in is supported only when the number of data nodes is greater than or equal to 2. In a single scale-in operation, the number of master nodes to be removed must be less than half of the number of master nodes before scale-in.
After scale-in, the disk usage must be less than 80%.
After scale-in, at least one instance of each node type must be retained in each AZ. For cross-AZ clusters, the difference in the number of nodes of the same type across different AZs must not exceed 1. For clusters with two AZs, at least two data nodes (ess) or cold data nodes (ess-cold) must be deployed in each AZ.

Calling Method

For details, see Calling APIs.

URI

POST /v1.0/{project_id}/clusters/{cluster_id}/node/offline

**Table 1** Path Parameters
Parameter	Mandatory	Type	Description
project_id	Yes	String	Definition Project ID. For details about how to obtain the project ID and name, see Obtaining the Project ID and Name. Constraints N/A Range Project ID of an account. The value contains 32 characters, consisting of lowercase letters and digits. Default Value N/A
cluster_id	Yes	String	Definition ID of the cluster to be scaled in. For details about how to obtain the cluster ID, see Obtaining the Cluster ID. Constraints N/A Range The value is a UUID containing 36 characters. Default Value N/A

Request Parameters

**Table 2** Request body parameters
Parameter	Mandatory	Type	Description
migrate_data	No	String	Definition Whether to migrate data from the node. Constraints N/A Range true: Migrate data. During data migration, the system moves all data shards from the node to be removed to the remaining nodes. The node is removed only after data migration is complete. false: Do not migrate data. When the data on the node to be removed has replicas on the remaining nodes, data migration can be skipped to shorten the cluster change time and reduce the load pressure caused by data migration. Default Value true
shrink_nodes	Yes	Array of strings	Definition ID of the node to be removed. Obtain the ID attribute in instances by referring to Querying Cluster Details. Constraints N/A
agency_name	No	String	Definition Agency name, which authorizes the current account to access and use OBS. OBS operation permissions are required to store snapshot data in the OBS bucket. Constraints The agency must have VPC operation permissions: "vpc:subnets:get","vpc:ports:". This parameter is mandatory when connected to the new IAM plane, and optional when connected to the old IAM plane. Range* The agency name can contain up to 64 characters. Only letters, digits, underscores (_), and hyphens (-) are allowed. Default Value N/A

Response Parameters

Status code: 200

Request succeeded.

None

Example Requests

Scale in a cluster by removing specified nodes.

POST https://{Endpoint}/v1.0/{project_id}/clusters/4f3deec3-efa8-4598-bf91-560aad1377a3/node/offline

{
  "shrink_nodes" : [ "2077bdf3-b90d-412e-b460-635b9b159c11" ],
  "migrate_data" : "true"
}

Example Responses

None

Status Codes

Status Code	Description
200	Request succeeded.
400	Invalid request. Modify the request before retry.
409	The request could not be completed due to a conflict with the current state of the resource. The resource that the client attempts to create already exists, or the update request fails to be processed because of a conflict.
412	The server did not meet one of the preconditions contained in the request.

Error Codes

See Error Codes.

Parent topic: Modify Configuration