Elasticsearch Agent Skill for intelligent cluster O&M

Skill overview

Install the skill

# Install or update Aliyun CLI
curl -fsSL https://aliyuncli.alicdn.com/setup.sh | bash

# Verify the version
aliyun version

# Enable automatic plugin installation
aliyun configure set --auto-plugin-install true
aliyun plugin update

Scenario 1: Create an Elasticsearch instance

The instance management skill lets you describe creation requirements in natural language. The AI assistant then automatically validates parameters and makes the API calls.

Prompt for the AI assistant:

In the China (Hangzhou) region, create a dual-availability zone Elasticsearch instance. Use version 8.17_with_X-Pack, 4 data nodes, each with a 20 GB ESSD, and the elasticsearch.sn2ne.large.new specification. Use VPC vpc-bp1***************, vSwitch vsw-bp1***************, and availability zone cn-hangzhou-i. Set the billing method to pay-as-you-go and name the instance log-es-cluster.

The AI assistant automatically performs the following actions:

1. Validates all required parameters, such as region, VPC, vSwitch, availability zone, password, and specification, prompting for any missing information.

2. Recognizes the dual-availability zone requirement and automatically configures zoneCount=2 and three dedicated master nodes.

3. Generates a clientToken (UUID) to ensure idempotence.

4. Calls the createInstance API to create the instance.

5. Queries the instance status to confirm successful creation.

Example command generated and executed by the AI assistant:

CLIENT_TOKEN=$(uuidgen)

aliyun elasticsearch create-instance \
  --region cn-hangzhou \
  --client-token $CLIENT_TOKEN \
  --es-admin-password "YourPassword@2024" \
  --es-version "8.17_with_X-Pack" \
  --node-amount 4 \
  --node-spec '{"disk":20,"diskType":"cloud_essd","spec":"elasticsearch.sn2ne.large.new"}' \
  --network-config '{"vpcId":"vpc-bp1***************","vswitchId":"vsw-bp1***************","vsArea":"cn-hangzhou-i","type":"vpc"}' \
  --payment-type postpaid \
  --description "log-es-cluster" \
  --zone-count 2 \
  --master-configuration '{"amount":3,"disk":20,"diskType":"cloud_essd","spec":"elasticsearch.sn2ne.xlarge"}' \
  --kibana-configuration '{"amount":1,"spec":"elasticsearch.n4.small"}' \
  --user-agent AlibabaCloud-Agent-Skills/alibabacloud-elasticsearch-instance-manage

Note the following constraints when you create an instance:

• Multi-availability zone deployment: When you set zoneCount to 2 or 3, you must configure three dedicated master nodes. For network configuration, you only need to provide the vswitchId for the primary availability zone. Nodes are automatically distributed across different availability zones.

• Idempotence protection: All write operations use a clientToken (in UUID format) to allow for safe retries after a request times out.

• Single-change principle: When scaling an instance, you cannot modify data nodes, master nodes, Kibana nodes, cold nodes, and coordinating nodes in the same API call.

• Specification differences: The supported node specifications vary by region. The skill includes a built-in specification map for automatic validation. In the China (Beijing), China (Shanghai), China (Hangzhou), and China (Shenzhen) regions, the specifications for new-generation cloud disk nodes require a .new suffix, for example, elasticsearch.sn2ne.large.new.

Scenario 2: Diagnose a cluster fault

The diagnosis skill uses a systematic four-step diagnostic workflow to address 8 major fault categories and incorporates 7 specialized Standard Operating Procedures (SOPs).

Prompt for the AI assistant:

My Elasticsearch cluster es-cn-xxx in the China (Hangzhou) region now has a Red health status. All production writes are failing. Urgently diagnose the cause and provide a solution.

The AI assistant automatically performs a four-step diagnosis:

Example diagnostic output:

P0 Critical Issue: Cluster health is Red
├─ Root Cause: Disk usage on node es-node-03 is at 96%, triggering the flood_stage read-only protection.
├─ Impact: Primary shards are unavailable, causing all write operations to fail.
├─ Recommended Actions:
│  1. Immediately delete expired indices to free up disk space.
│  2. After disk usage drops below 85%, run the following command:
│     PUT _all/_settings {"index.blocks.read_only_allow_delete": null}
│  3. Evaluate whether to scale up disk capacity.
└─ Reference SOPs: sop-disk-storage.md / sop-cluster-health.md

PUT _all/_settings
{
  "index.blocks.read_only_allow_delete": null
}

Scenario 3: Configure network security

The network management skill automatically identifies the instance's architecture type (cloud-native or basic-managed) and selects the appropriate API for network configuration.

Prompt for the AI assistant:

Enable HTTPS for the Elasticsearch instance es-cn-xxx. Also, append the IP ranges 203.0.113.0/24 and 198.51.100.0/24 to the public whitelist without overwriting the existing entries.

The AI assistant automatically performs the following actions:

1. Calls DescribeInstance to check the instance status and current HTTPS state.

2. If HTTPS is not enabled, it calls OpenHttps to enable it.

3. Calls ModifyWhiteIps using the Append mode to add the IP addresses to the whitelist.

# Check the current HTTPS status
aliyun elasticsearch describe-instance \
  --instance-id es-cn-xxx \
  --user-agent AlibabaCloud-Agent-Skills/alibabacloud-elasticsearch-network-manage

# Enable HTTPS
aliyun elasticsearch open-https \
  --instance-id es-cn-xxx \
  --user-agent AlibabaCloud-Agent-Skills/alibabacloud-elasticsearch-network-manage

# Append to the whitelist (Append mode)
aliyun elasticsearch modify-white-ips \
  --instance-id es-cn-xxx \
  --modify-mode Append \
  --white-ip-group '{"groupName":"default","ips":["203.0.113.0/24","198.51.100.0/24"],"whiteIpType":"PUBLIC_ES"}' \
  --user-agent AlibabaCloud-Agent-Skills/alibabacloud-elasticsearch-network-manage

Key constraints for network configuration:

Related links

• Alibaba Cloud Agent Skills Portal

• Alibaba Cloud Elasticsearch Product Page

• Elasticsearch API Reference