Create an Alibaba Cloud Elasticsearch cluster

Billing

The prices and billable items of Alibaba Cloud ES are described in billing rules.

Procedure

1. Log on to the Alibaba Cloud Elasticsearch console.

2. In the left navigation menu, choose Elasticsearch Clusters.

3. On the Elasticsearch Clusters page, click Create in the upper-left corner.

   Parameter	Description
   Product Type	Subscription: Recommended when you use an ES cluster for a long period. You pay once, and the price is lower than that of pay-as-you-go. Pay-as-you-go: Suited to short-term scenarios such as program development and feature testing. You are billed by the hour and can release the cluster at any time to stop the charges. You can create a pay-as-you-go cluster first and switch it to subscription after it meets your requirements.
   Regions and Zones	Select a region based on your business needs. The region cannot be changed after the cluster is created: The closer the region is to your users, the lower the network latency and the faster the access speed. If the ES cluster must communicate with other Alibaba Cloud services over an internal network, deploy them in the same region.
   Number of Zones	1-AZ: A standard deployment mode suited to non-critical workloads. 2-AZ: A cross-zone disaster recovery deployment mode suited to production workloads. 3-AZ (default): A high availability (HA) deployment mode suited to production workloads that have higher availability requirements. After you select two or three zones, the page displays the Zone 2 and Zone 3 configuration items. The default value is Random, which lets the system assign the zones automatically. You can also specify them manually. For more information, see Deploy and use a multi-zone cluster.
   Network Type	Only Virtual Private Cloud (VPC) is supported.
   VPC	Select a VPC in the selected region: If the ES cluster must communicate with other Alibaba Cloud services over an internal network, deploy them in the same VPC. If no VPC meets your business needs, create one first by following Create a VPC and a vSwitch.
   Virtual Switch	Only the vSwitches that reside in the specified VPC and in the same zone as the ES cluster are displayed. If no vSwitch meets your business needs, create one first by following Create a VPC and a vSwitch.
   Cluster Type	Engine-enhanced edition: 100% compatible with open-source Elasticsearch and provides enterprise features such as X-Pack free of charge. It is optimized for search, logging, and AI scenarios, and you can enable its enhanced capabilities on demand. The cloud-native control mode improves change efficiency, and the intelligent O&M ES Agent is ready to use out of the box. Standard Edition: 100% compatible with open-source Elasticsearch and provides enterprise features such as X-Pack free of charge. It covers multiple ES versions from 5.x to 9.x. The cloud-native control mode improves change efficiency, and the intelligent O&M ES Agent is ready to use out of the box. For more information, see Version features comparison and Feature overview.
   Elasticsearch Version	Engine-enhanced edition: Supports versions 9.4.0, 8.17.0, and 7.10.0, and the 9.x (AI Engine) version. Click Other to select version 8.15.1, 7.16.2, or 6.7.0. Standard Edition: Supports versions 9.3.2, 8.13.4, 8.9.1, 8.5.1, 7.7.1, 6.8.23, and 5.6.16. Compare the supported features of each version in Version features comparison and Release Note. The regions and zones that each version supports may differ. The purchase page displays the available options.
   Scenario selection and advanced features	Select a scenario based on your business type. The available scenarios are determined by the Elasticsearch Version that you select. After you select a scenario, the page displays the advanced enhanced features that apply to the scenario and their billing methods. You can enable these features on demand: Search scenario: Applies to common scenarios that require high query performance. FalconSeek high-performance search engine: Built on the Havenask search engine technology that Alibaba Group developed in house, and deeply enhances the Elasticsearch kernel. It is fully compatible with Elasticsearch syntax and ecosystem while delivering stronger search performance and higher cost-effectiveness. Billing method: free to enable, with no additional billable items. OpenStore storage engine (high-performance search): Based on high-performance shared storage and intelligent caching to ensure low-latency retrieval. Scaling does not require a full migration of the underlying data, which significantly shortens the duration of cluster changes. Scheduled elasticity and on-demand elasticity allow compute resources to scale flexibly with your business peaks and valleys. Billing method: OpenStore storage space (DFS) plus cache disk space. Logging scenario: Applies to workloads that write more than they read, hold time-series data, and require long-term data storage and analysis but have low real-time requirements. Indexing Service managed write service: A cloud-managed service that delivers more than 10x elastic write throughput. You do not need to reserve resources for peak write traffic, and you are billed for the actual write traffic, which enables low-cost analysis of massive logs. Billing method: write traffic and managed storage space. OpenStore storage engine (log analytics): Based on object storage and intelligent caching, which significantly reduces the cost of retaining massive logs over the long term. You do not need to reserve capacity in advance, and historical data can be queried without restoration. Billing method: OpenStore storage space (OSS), with storage as low as CNY 0.15 per GB per month. AI scenario: This is the only scenario supported when you select the 9.x (Agent Engine Edition) version. In this case, you cannot select the search scenario or the logging scenario. SliceCollection massive-scale multi-tenant search: Designed for agent retrieval and RAG workloads. A single cluster can host tenant data on the scale of hundreds of millions, which suits scenarios such as knowledge bases and code retrieval. Billing method: based on shared storage that implements storage-compute disaggregation and read/write splitting. Write resources and query resources scale independently, and dormant data incurs only storage fees. Engine-enhanced edition and Standard Edition both require a scenario. Engine-enhanced edition provides the advanced enhanced features described above, whereas Standard Edition requires a scenario but does not provide advanced enhanced features. After the cluster is created, you can modify the scenario configuration as needed to optimize the cluster and index configurations and to reduce the cluster exceptions and performance issues that improper configurations cause, as described in Modify a scenario-specific configuration template. The functions and features that each version supports differ. For details, see Feature overview.
   Deployment Mode	cloud-native new control mode: The deployment mode that the current purchase page provides. It supports a wide range of cluster types and versions, provides out-of-the-box AI capabilities, and significantly improves cluster O&M efficiency in areas such as scaling, monitoring, and alerting.
   Specifications	Click each node section and configure the nodes based on your business needs. The configurable nodes are data node, Kibana node, dedicated master node, cold data node, client node, and frozen data node. The node specification family of the data nodes is determined by the selected scenario: the search scenario uses the next-generation cloud disk type, and the logging scenario uses the storage-compute disaggregated type. After the cluster is created, you cannot change the node specification family, and you cannot scale node specifications across families. Select node specifications based on the guidance in Capacity assessment. Configure each node type as described in Node configurations for an Elasticsearch cluster. Click Reset to Default in the lower-right corner of the section to restore the modified node configurations to the default configurations.
   Resource Group	Resource groups let you manage ES clusters in groups. For details, see Resource groups and authorization.
   Cluster Name	If you leave this parameter empty, the system generates a name automatically. You can change the cluster name on the basic information page after the cluster is created.
   Username	The default account is elastic and cannot be changed. This account holds the super administrator (Superuser) role of the cluster and is used to access the ES cluster and to log on to the Kibana console. Important A change to the password of the elastic account takes effect immediately and interrupts all existing connections that use this account because of authentication failures. If the password is hard-coded in your business applications, the service may become unavailable. For security reasons, the principle of least privilege prohibits using the elastic account directly in the code or applications of your production business. Create roles and common users that have specific permission scopes in the Kibana console for business use, as described in Manage user permissions with X-Pack roles.
   Password	The password of the elastic account. This parameter cannot be empty.
   Subscription Duration	You must select a subscription duration when the product type is subscription. We recommend that you enable Auto-renewal to prevent the cluster from being stopped or released after the resources expire. If auto-renewal is enabled for a monthly subscription, the Unified Auto Renewal Cycle is one month. If auto-renewal is enabled for a yearly subscription, the Unified Auto Renewal Cycle is one year. The selectable durations are one month, one year, two years, three years, and five years.

4. Click Buy Now to go to the Confirm Order page. Confirm that the cluster configuration is correct, and then click Activate.

   Note

   In pay-as-you-go mode, the button text is Create Now.

5. On the Elasticsearch Clusters page, select the region where you created the cluster in the top menu bar, and then view the created ES cluster in the cluster list.

   After the cluster is created, it takes some time to take effect. The duration depends on the cluster specifications, the data structure, and the data size. You can view the instance status in the Status column.

What to do next

• View the basic information about the cluster, such as the private endpoint, the running status, and the node information, in View cluster status and node information.

• Access the ES cluster in one of the ways described in Connect to a cluster.

• Migrate data from other services or collect data into ES as described in Best practices overview.

• If your application requires an account and a password for the ES cluster, create roles and users that have the required permissions in the Kibana console, as described in Manage user permissions with X-Pack roles.

• Modify a scenario-specific configuration template to optimize the cluster and index configurations and to reduce the cluster exceptions and performance issues that improper configurations cause, as described in Modify a scenario-specific configuration template.

• After you purchase cold data nodes, use index lifecycle management to maintain hot and cold data automatically and separate hot data from cold data, as described in Use index lifecycle management to manage Heartbeat data.

• After the cluster is created, scale the cluster up or down to adjust configurations such as the node specifications and the capacity, as described in Scale up a cluster and Scale down a cluster.

FAQ

• Can I purchase a single-node Elasticsearch instance?

• What should I do if resources are sold out when I try to create an instance?

• Alibaba Cloud Elasticsearch cluster FAQ