What's New

1

Optimized user guide structure

To improve user experience, the CSS User Guide has been optimized as follows:

• Simplified titles: Optimized the title structure, and removed redundant information from sub-section titles for faster information retrieval.

• Redesigned usage procedures: Added new flowcharts and optimized procedures for using Elasticsearch, OpenSearch, and Logstash.

• Added network configuration guidelines: a detailed introduction to four different networking options.

• Split the log management section: There are now three independent topics – querying logs, ingesting logs into a cluster, and backing up logs to OBS.

• Added plugin management: Added a new section about default cluster plugins.

These optimizations make the document structure clearer and content more accessible and helpful.

Commercial use

CSS User Guide

2

Independent documentation and optimized content for the vector database

To align with the product strategy, the vector database documentation has been separated from the Elasticsearch and OpenSearch documentation to form an independent module. A guide comparing the CSS vector engine with the open-source OpenSearch Vector Engine has been added, clearly explaining the differences between the two and how to choose between them based on your service requirements. A quick start guide to OpenSearch Vector Engine is also provided, with step-by-step procedures to ease the learning curve.

The CSS vector engine documentation has been significantly enriched, including a redesigned usage process, a better description of the underlying technology, clearer explanations of constraints, and additional guidelines on applications scenarios.

These optimizations focus on technical depth and higher operational efficiency, helping developers quickly build high-performance vector search applications.

Commercial use

CSS Vector Engine vs. OpenSearch Vector Engine

1

Optimized documentation on CSS enhancements over community versions

Documentation on CSS enhancements over community versions has been optimized as follows:

• Background and principles: Added descriptions of technical principles, service scenarios, and cluster performance impact.

• Constraints: Updated descriptions of constraints and restrictions.

• Parameter description: Unified the parameter description style and enriched parameter descriptions to make them easier to understand.

• Content structure: Optimized the structure of configuration operation guidelines to make them easier to follow and better match real-world scenarios.

• Reminder style: Incorporated redundant notes, reminders, and warnings into the body text to improve readability.

These optimizations help users better understand product features and improve configuration efficiency by providing more detailed descriptions of technical principles and easier-to-follow instructions and procedures.

Commercial use

CSS enhancements over open-source Elasticsearch

CSS enhancements over open-source OpenSearch

1

Optimized best practices documentation

To accommodate growing case studies, the Best Practices documentation has been optimized to improve accessibility:

Reconstructed content structure: Case studies have been reclassified based on competitive features and ranked by importance.

Scenario realignment: The document structure is now aligned with real-world scenarios to improve accessibility.

These optimizations ensure that the best practices better reflect real-world scenarios, helping users quickly locate relevant best practices and solutions to their problems.

Commercial use

CSS Best Practices

1

Intelligent O&M upgrade

To overcome the limitations of traditional O&M methods, CSS introduces an upgraded Intelligent O&M feature for comprehensive, automated cluster health checks and management.

• Automatic diagnostics: Supports both scheduled and manual diagnostics, covering over 10 key metrics, such as data node disk usage, cluster health status, and node CPU usage.

• Enhanced operational efficiency: Through automated reporting and intelligent troubleshooting suggestions, manual diagnosis is accelerated, and cluster reliability and availability are enhanced.

Commercial use

Intelligent Risk Detection for Elasticsearch Clusters

Intelligent Risk Detection for OpenSearch Clusters

2

Added compatibility for OpenSearch 2.19.0

OpenSearch clusters can be seamlessly upgraded to 2.19.0 across versions. This version integrates kernel-level enhancements from CSS on top of the open-source OpenSearch 2.19.0. These enhancements include improved ingestion performance, optimized aggregation performance, read/write splitting, and large query isolation.

Commercial use

Upgrading the Version of an OpenSearch Cluster

3

Adding dedicated cold data nodes

To optimize storage costs and query performance in the face of massive data volumes, CSS allows you to add dedicated cold data nodes for separate management of hot and cold data.

• Dynamic expansion: Cold data nodes can be added on demand to adapt to real-time data growth.

• Storage optimization: Cold data nodes use less expensive hardware (such as HDDs), significantly reducing storage costs.

• Isolation: Tier storage management ensures that cold data won't affect the query performance of hot data.

• Convenient management: Cold data node policies can be conveniently configured on the console to automate data migration and lifecycle management.

Commercial use

Scaling Out/Up an Elasticsearch Cluster

Scaling Out/Up an OpenSearch Cluster

4

Compatibility with the latest Elasticsearch client versions

To ensure compatibility with the mainstream Elasticsearch ecosystem, CSS supports the High Level Rest Client (elasticsearch-rest-high-level-client) that comes with Elasticsearch 7.x and 8.x.

• Version adaptation: Supports the APIs of Elasticsearch 7.x and 8.x to ensure the continuity of enterprise applications.

• Stability assurance: Thorough compatibility testing ensures smooth client-cluster collaboration.

Commercial use

Connecting to an Elasticsearch Cluster Through the Elasticsearch Java API Client

Connecting to an OpenSearch Cluster Through the Elasticsearch Java API Client

1

Hosted OpenSearch 2.19 service

CSS has been upgraded to the OpenSearch 2.19 kernel. This version provides a comprehensive, enterprise-grade suite of open-source search and analytics services, creating a sustainable technical foundation. It delivers enhanced management and kernel capabilities, offering high performance, availability, and cost-effectiveness.

OpenSearch is a community-driven, open-source fork of Elasticsearch, initially led by AWS. It is a 100% open-source search and analytics suite used for a broad set of use cases. While retaining all the core functionalities of Elasticsearch, OpenSearch differentiates itself through its commitment to open governance and accessibility. Key advantages include:

• True open source: The Apache 2.0 open-source license guarantees unlimited commercial use.

• Zero-barrier access to advanced capabilities: Enterprise-grade features, such as the security engine, real-time alerting, SQL query, and machine learning, are all available free of charge.

• Transparent community governance: The project's roadmap and development are guided by a diverse community, which prevents vendor lock-in and ensures the platform evolves to meet user needs and emerging technology trends.

Commercial use

Using OpenSearch for Data Search

1

Added support for changing a cluster's subnet

If the current subnet of an existing Elasticsearch or OpenSearch cluster cannot meet your requirements, you can switch to another subnet.

After switching the subnet, the new subnet is used for adding nodes, including scaling, adding dedicated master/client nodes, and enabling VPC Endpoint. IP addresses will be allocated to newly added nodes from the new subnet. This operation will not alter the network settings of existing nodes.

Commercial use

Can I Expand the Subnet for an Elasticsearch or OpenSearch Cluster?

2

Added the new cluster creation page

To improve efficiency in creating clusters, CSS has made several enhancements on the cluster creation page in terms of ease-of-use. Being more intuitive, the new UI streamlines cluster creation procedures.

Commercial use

Creating an Elasticsearch Cluster (New Version)

Creating an OpenSearch Cluster (New Version)

Creating a Logstash Cluster (New Version)

3

Added support for configuring cluster routes

By default, a CSS Elasticsearch or OpenSearch cluster cannot access a target service—such as a client program or a Lightweight Directory Access Protocol (LDAP) service—that is located in a different VPC. To enable it, you need to configure a route for the cluster.

Commercial use

Configuring Routes for an Elasticsearch Cluster

Configuring a Route for an OpenSearch Cluster

4

Added monitoring of write failures and primary/secondary synchronization jobs

Metrics that monitor write failures for Elasticsearch and OpenSearch clusters and the primary/secondary synchronization jobs for Elasticsearch clusters have been added. They help quickly identify and locate issues and improve diagnostic and troubleshooting efficiency.

Commercial use

Monitoring Metrics for Elasticsearch Clusters in Cloud Eye

Monitoring Metrics for OpenSearch Clusters in Cloud Eye

1

SQ4/SQ8 support by vector DB

With Elasticsearch 7.10.2 clusters, vector search supports the GRAPH_SQ4 and GRAPH_SQ8 indexing algorithms. Combining the HNSW algorithm with the SQ algorithm, these algorithms reduce the storage overhead of the original vectors and improves build and query efficiency.

Commercial use

Creating Vector Indexes in an Elasticsearch Cluster

2

Public network access for Logstash clusters

An NAT gateway and return routes need to be configured to enable Logstash clusters to access data sources on the Internet, ensuring proper data ingestion, storage, and processing across different networks.

Commercial use

Configuring Public Network Access for a Logstash Cluster

3

Support for OpenSearch 2.17.1

With CSS, you can create OpenSearch 2.17.1 clusters. They are fully compatible with open-source OpenSearch 2.17.0 clusters. For details about open-source OpenSearch, see OpenSearch and OpenSearch Dashboards 2.17.0 Release Notes.

Commercial use

OpenSearch Cluster Planning Suggestions

1

New case study: Migrating Kibana Saved Objects Between Elasticsearch Clusters

Describes procedures for migrating Kibana's saved objects between different Elasticsearch clusters.

Commercial use

Migrating Kibana Saved Objects Between Elasticsearch Clusters

2

Elasticsearch supports vector search for nested fields

In Elasticsearch clusters, nested fields allow multiple vectorized records to be stored in a single document. For a document that contains multiple vector records, if the query vector matches any of them, the document is returned.

Commercial use

Using a Vector Index with Nested Fields

3

New case study: Synchronizing Data from RDS for MySQL to Elasticsearch Through Logstash

By configuring the Logstash configuration file to define the JDBC input and Elasticsearch output, you can synchronize data from RDS for MySQL to Elasticsearch at specified intervals.

Commercial use

Synchronizing Data from RDS for MySQL to Elasticsearch Through Logstash

4

Synchronizing Elasticsearch Alerts to Prometheus

You can synchronize Elasticsearch alerts to Prometheus so you can use Prometheus to monitor the key performance metrics of Elasticsearch clusters in real time.

Commercial use

Synchronizing Elasticsearch Alerts to Prometheus

5

Elasticsearch supports offline creation of vector indexes

Where real-time data is not crucial, you can enable offline index creation to reduce the repeated creation of intermediate vector indexes, improving end-to-end write and index merging performance.

Commercial use

Creating Vector Indexes in an Elasticsearch Cluster

6

New case study: Testing the Performance of CSS's Elasticsearch Vector Search

Describes how to use open-source datasets and pressure testing tools to test the performance of CSS's Elasticsearch vector search engine.

Commercial use

Testing the Performance of CSS's Elasticsearch Vector Search

1

Support for pre-filtering query by Elasticsearch

Elasticsearch clusters support pre-filtering queries on vector indexes. Pre-filtering query works like this: First, filters are applied to indexes to generate an intermediate result set. Then, vector similarity-based search is performed on these results to find the most relevant vectors. This improves search efficiency and accuracy.

Commercial use

Using Vector Indexes for Data Search in an Elasticsearch Cluster

2

Archive storage support by Elasticsearch clusters that have enabled storage-compute decoupling

Elasticsearch clusters with decoupled storage and compute support the change of data storage classes in OBS. Changing the storage class from standard to archive in OBS helps cut the cost of storing historical data.

Commercial use

Configuring Storage-Compute Decoupling for an Elasticsearch Cluster

Decoupling Index Storage and Compute in an Elasticsearch Cluster Through Index Lifecycle Management

3

Added best practices for Elasticsearch data migration

Added best practices for migrating data between Elasticsearch clusters using Reindex API and Elasticsearch Migration Tool (ESM).

Commercial use

Migrating Data Between Elasticsearch Clusters Using the Reindex API

Migrating Data Between Elasticsearch Clusters Using ESM

4

Added monitoring metrics for vector indexes

Metrics that monitor the off-heap memory usage/utilization of vector indexes are added for Elasticsearch and OpenSearch clusters. They help quickly identify and locate issues and improve diagnosis efficiency.

Commercial use

Monitoring Metrics for Elasticsearch Clusters in Cloud Eye

Monitoring Metrics for OpenSearch Clusters in Cloud Eye

5

Added description about the performance impact of Elasticsearch data migration

The impact of data migration on the performance of the source Elasticsearch cluster is described, helping users plan the data before and after the migration to minimize this impact and ensure service continuity.

Commercial use

Migrating Data Between Elasticsearch Clusters Using Huawei Cloud Logstash

1

Support for vector search by OpenSearch

Added the operation guide and code examples for using vector search with an OpenSearch cluster.

Commercial use

Configuring Vector Search for OpenSearch Clusters

2

Index lifecycle management for OpenSearch clusters

For an OpenSearch cluster, index lifecycle policies can be configured to decouple storage and computation for index data.

Commercial use

Decoupling Index Storage and Compute Through Index Lifecycle Management

3

Node-level disk usage monitoring

The disk usage metric has been added for Elasticsearch and OpenSearch clusters.

Commercial use

Monitoring Metrics for Elasticsearch Clusters in Cloud Eye

Monitoring Metrics for OpenSearch Clusters in Cloud Eye

4

Load balancers and VPC Endpoint Service support IPv6 networks

Elasticsearch and OpenSearch clusters in CSS can be accessed via IPv6 addresses through a dedicated load balancer or VPC endpoint.

Commercial use

Configuring VPC Endpoint Service for an Elasticsearch Cluster

Configuring a Dedicated Load Balancer for an Elasticsearch Cluster

Configuring VPC Endpoint Service for an OpenSearch Cluster

Configuring a Dedicated Load Balancer for an OpenSearch Cluster

5

CSS Dashboard displays monitoring metrics

On the Dashboard page of the CSS console, you can check monitoring metrics and configure alarm rules.

Commercial use

CSS Resource Monitoring

6

Adapting CSS to the latest CTS specifications

Updated the list of key operations that can be audited through CTS based on the latest CTS specifications.

Commercial use

Viewing Elasticsearch Cluster Audit Logs

Viewing OpenSearch Cluster Audit Logs

Viewing Logstash Cluster Audit Logs

7

Disk I/O monitoring

Eight metrics for I/O monitoring have been added for Elasticsearch and OpenSearch clusters.

Commercial use

Monitoring Metrics for Elasticsearch Clusters in Cloud Eye

Monitoring Metrics for OpenSearch Clusters in Cloud Eye

8

Read/write splitting enables automatic metadata synchronization

The read/write splitting function enables automatic metadata synchronization for Elasticsearch clusters.

Commercial use

Configuring Read/Write Splitting Between Two Elasticsearch Clusters

9

Disk type change

Change the node storage type of an Elasticsearch or OpenSearch cluster by node type. To do that, you need to delete old disks when rebuilding a node, and then attach disks of the new type.

Commercial use

Changing the Node Specifications of an Elasticsearch Cluster

Changing the Node Specifications of an OpenSearch Cluster

10

Accessing clusters through LDAP

Enable Lightweight Directory Access Protocol (LDAP) authentication for an Elasticsearch or OpenSearch cluster in CSS to allow access by LDAP users of specific roles.

Commercial use

Accessing an Elasticsearch Cluster Using LDAP

Accessing an OpenSearch Cluster Using LDAP

11

Added performance testing data for enhanced aggregation

Added performance testing data for the enhanced aggregation function for Elasticsearch clusters.

Commercial use

Configuring Enhanced Aggregation for an Elasticsearch Cluster

12

Modified the tag details page

The cluster tag details page is redesigned, so the operation guide for tag management is updated accordingly.

Commercial use

Setting Tags for an Elasticsearch Cluster

Setting Tags for an OpenSearch Cluster

Setting Tags for a Logstash Cluster

13

Minimizing permissions for IAM agencies

Permissions can be minimized for IAM agencies automatically created for CSS. If there is an IAM agency automatically created earlier, you can also click One-click authorization to minimize its permissions.

Commercial use

All operation documents that contain steps about agency configuration

1

Optimization of Best Practices

Made the documents easier to understand and more searchable.

• Described best practices in terms of application scenario, solution architecture, and advantages.

• Enriched best practices on cluster migration.

Commercial use

Elasticsearch Cluster Migration

2

Optimized FAQs

Made the documents easier to understand and more searchable.

• Optimized topic titles to describe issues from the perspective of users.

• Retired questions that are no longer relevant and rarely visited.

Commercial use

FAQs

3

Updated User Guide

Made the documents easier to understand and more searchable.

• Rearranged the table of contents to make information more intuitive and searchable.

• Included scenario names in topic titles so that users can find needed information more quickly.

• Put documents for the same task together to offer all information needed.

• Added application scenarios and use constraints in operation guides.

Commercial use

Procedure for Using Elasticsearch

Procedure for Using OpenSearch

Procedure for Using Logstash

1

OBS cold storage monitoring in the case of storage-compute decoupling

When storage-compute decoupling is enabled, the real-time OBS rates are recorded in the system index.

Commercial use

Configuring Storage-Compute Decoupling for an Elasticsearch Cluster

2

Optimized the Getting Started guide

Made the documents easier to understand and more searchable.

• Added scenario-specific cases for OpenSearch and Logstash clusters to help beginners get started quickly.

• Optimized topic titles to better identify specific tasks.

Commercial use

Using OpenSearch for Data Search

Using Logstash for Data Migration

1

Pipeline monitoring for Logstash clusters

Logstash clusters in CSS support pipeline-level event monitoring. Alarm can be customized based on pipeline monitoring metrics.

Commercial use

Monitoring Metrics for Logstash Clusters in Cloud Eye

2

Showing the causes of log backup failures

On the log backup page of CSS, information about failed log backup tasks and failure causes are displayed, facilitating troubleshooting.

Commercial use

Querying and Managing Elasticsearch Cluster Logs

Querying and Managing OpenSearch Cluster Logs

Querying and Managing Logstash Cluster Logs

3

Index recycle bin for Elasticsearch clusters

To allow users to restore data after misdeletion, CSS provides an index recycle bin for Elasticsearch clusters. Similar to other recycle bins, the index recycle bin temporarily stores deleted indexes so that users can restore them before they are finally removed from the recycle bin. This helps to improve cluster data reliability.

Commercial use

Configuring an Index Recycle Bin for an Elasticsearch Cluster

4

Real-time log collection

The real-time logs of an Elasticsearch or OpenSearch cluster can be stored within that cluster, facilitating log search and analysis using Kibana.

Commercial use

Querying and Managing Elasticsearch Cluster Logs

Querying and Managing OpenSearch Cluster Logs

5

Modifying preset word dictionaries

The preset Static Main Word Dictionary, Static Stop Word Dictionary, Extra Main Word Dictionary, and Extra Stop Word Dictionary can be added, deleted, and modified for Elasticsearch and OpenSearch clusters.

Commercial use

Configuring Custom Word Dictionaries for an Elasticsearch Cluster

Configuring Custom Word Dictionaries for an OpenSearch Cluster

1

Cross-engine upgrade of Elasticsearch clusters

Elasticsearch clusters of version 7.10.2 can be upgraded to OpenSearch clusters of version 1.3.6.

Commercial use

Upgrading the Cluster Version

2

Cluster status check during cluster specification modification

When the specifications of a cluster are changed, the cluster status can be checked. The specifications of nodes are changed one by one to ensure success and data security.

Commercial use

Changing the Node Specifications

1

Support for general AC computing-plus ECSs

CSS supports general AC computing-plus ECSs equipped with scalable processors, providing enhanced computing, security, stability, and high cost-effectiveness.

Commercial use

Cloud Search Service Price Calculator

1

Opensearch Alert Plugin

By default, CSS has installed the open-source alert plugin opensearch-alerting for OpenSearch clusters to send notifications when data meets specific conditions.

Commercial use

Using the OpenSearch Alert Plugin

2

Selecting a shared VPC during cluster creation

A VPC is required when you create a cluster. You can select the name and ID of the VPC shared with the current account.

Commercial use

Creating an Elasticsearch Cluster

Creating a Logstash Cluster

Creating an OpenSearch Cluster

1

OpenSearch clusters

You can create and manage OpenSearch clusters in CSS.

Commercial use

Creating an OpenSearch Cluster

2

Auto-renewal for yearly/monthly clusters

For yearly/monthly clusters with auto-renewal enabled, you can modify the auto-renewal function.

Commercial use

Modifying Auto-Renewal

1

Batch export of Logstash cluster configuration files

The configuration files of Logstash clusters can be exported in batches.

Commercial use

Exporting Configuration Files

1

Extreme SSD storage

Elasticsearch and Logstash clusters support the extreme SSD storage.

Commercial use

Creating an Elasticsearch Cluster

Creating a Logstash Cluster

2

Migrating data before scaling in or replacing a specified node

You can select the Whether to perform data migration option to migrate data. If the target node have disabled indexes or indexes that have no replicas, this option must be selected.

Commercial use

Replacing Specified Nodes for an Elasticsearch Cluster

Scaling In an Elasticsearch Cluster

3

Logstash cluster force restart

If a Logstash cluster is faulty due to long-term running or other unknown reasons, you can forcibly restart the cluster to restore it.

Commercial use

Forcibly Restarting VMs in a Cluster

4

Querying clusters by enterprise project or private IP address

In the cluster list, you can quickly query clusters by enterprise project and private IP address.

Commercial use

Viewing Elasticsearch Cluster Information

5

Checking cluster load when upgrading an Elasticsearch cluster

If a cluster is overloaded, the upgrade task may suspend or fail. Cluster load check can effectively prevent upgrade task failures.

Commercial use

Creating an Upgrade Task

1

Binding load balancers on the CSS management console

When interconnecting with a dedicated load balancer, you can bind a load balancer and add a listener on the CSS management console.

Commercial use

Connecting to a Dedicated Load Balancer

2

Intelligent O&M

CSS provides intelligent O&M to detect potential cluster risks and provide risk handling suggestions.

Commercial use

Intelligent Risk Detection for Elasticsearch Clusters

3

Upgrading the service to version 6.8.23

CSS clusters of the 6.2.3 and 6.5.4 versions can be upgraded to the version 6.8.23.

Commercial use

Upgrading Versions

4

Logstash cluster

CSS supports Logstash clusters. You can modify the Logstash configuration files in the configuration center to import multiple data sources, such as edis, elasticsearch, and jdbc, to Elasticsearch clusters.

Commercial use

Logstash

5

Read/Write splitting

CSS supports read/write splitting. Data written to the primary cluster (Leader) can be automatically synchronized to the secondary cluster (Follower). In this way, data is written to the primary cluster and queried in the secondary cluster.

Commercial use

Read/Write Splitting

6

Enhanced import performance

Developed by CSS, this feature effectively improves import performance and reduces write rejection by optimizing bulk routing, text index acceleration, and word segmentation acceleration.

Commercial use

Enhanced Import Performance

7

Enhanced aggregation

The enhanced aggregation is an optimization feature for service awareness. With this feature, you can optimize the aggregation analysis capability of observable services.

Commercial use

Enhanced Aggregation

1

Supported the open-source OpenDistro alarm plug-in

By default, the open-source OpenDistro alarm plugin (opendistro_alerting) is integrated into CSS to send notifications when data meets specific conditions.

Commercial use

Using the OpenDistro Alarm Plugin

2

Supported switching between AZs

CSS supports cross-AZ deployment. You can add an AZ to obtain more resources or improve cluster availability, and can migrate your current AZ to one with higher specifications.

Commercial use

Changing AZs

3

Version upgrading

Same version upgrade and cross-version upgrade are supported. Same version upgrade is to upgrade the kernel patch of a cluster to fix problems or optimize performance. Cross-version upgrade is to upgrade the cluster version to enhance functions or incorporate versions.

Commercial use

Upgrading Versions

4

Node monitoring

You can query the monitoring metrics and alarms generated for CSS nodes on the Cloud Eye console or using CSS APIs.

Commercial use

Monitoring Metrics

1

OBS Operations Monitoring

The real-time OBS rate metric is added to CSS and recorded in the system index. You can view the usage of OBS in real time.

Commercial use

Monitoring OBS Operations

2

Yearly/Monthly billing mode

CSS clusters can be purchased in yearly/monthly mode. This mode is recommended for long-term users.

Commercial use

Yearly/Monthly

3

Decreasing Cluster Specifications

You can decrease cluster specifications. After decreasing cluster specifications, the cluster performance will deteriorate and service capabilities will be affected. Exercise caution when performing this operation.

Commercial use

Changing the Node Specifications

1

Replacing a Specified Node

If a node in the cluster is faulty, you can create a new node with the same specifications to replace it.

Commercial use

Replacing a Specified Node

2

Improved Cold Data Query Performance

To accelerate data queries for the first time on the Discover page of Kibana, you can cache data locally.

Commercial use

Enhanced Cold Data Query Performance

3

Independent Master/Client Nodes

If the workloads on the data plane of a cluster increase, you can dynamically scale the cluster by adding master or client nodes.

Commercial use

Adding Master/Client Nodes

4

Changing the Security Mode in One-click

After a cluster is created, you can change its security mode in one-click.

Commercial use

Changing the Security Mode

5

Switching the Security Group of Elasticsearch Cluster

After an Elasticsearch cluster is created, you can modify its security group.

Commercial use

Viewing Cluster Information

6

Added Traffic Control 2.0

After the client write traffic backpressure and control is enabled, large requests will be rejected when too much node heap memory is occupied. This function prevents nodes from breaking down and reduces the risk of node unavailability.

Commercial use

Flow Control 2.0

7

Connecting a Cluster to a Dedicated Load Balancer

You can connect a cluster to a dedicated load balancer. CSS integrates shared load balancers and allows you to bind public network access and enable the VPC Endpoint service. Dedicated load balancers provide more functions and higher performance than shared load balancers.

Commercial use

Connecting a Cluster to a Dedicated Load Balancer

1

Best Practices

This document summarizes common best practices for CSS. Operation details are described for each practice.

Commercial use

Best Practices

1

HTTP Status Code Monitoring

CSS allows you to monitor the HTTP status codes of clusters so that you can accurately view the actual status of calling Elasticsearch APIs.

Commercial use

HTTP Status Code Monitoring

2

P99 Latency Monitoring

CSS allows you to monitor the P99 latency of search requests in clusters to effectively reflect the actual search performance.

Commercial use

P99 Latency Monitoring

1

Vector retrieval

CSS supports k-nearest neighbor (KNN) and approximate nearest neighbor (ANN) vector search.

Commercial use

Feature Description

1

Storage-compute decoupling

CSS allows you to store historical data in OBS to reduce data storage costs.

Commercial use

Feature Description