Learn about the feature updates and documentation for each Alibaba Cloud E-MapReduce release.
For more details, see release notes.
2025
October
Feature | Description | Release date | Documentation |
EMR-5.21 & EMR-3.55 released | This release optimizes components such as Hive, Spark, Tez, and Ranger, and fixes issues in previous versions. | 2025-10-27 |
July
Feature | Description | Release date | Documentation |
EMR-5.20 & EMR-3.54 released | This release fixes issues in previous versions. | 2025-07-10 |
April
Feature | Description | Release date | Documentation |
EMR-5.19 & EMR-3.53 released | This release upgrades multiple service versions and includes comprehensive fixes and optimizations for issues in previous versions. | 2025-04-24 |
2024
December
Feature | Description | Release date | Related documentation |
EMR 5.18.1 & EMR 3.52.1 release | This release upgrades multiple services, fixes issues from previous versions, and discontinues services such as Impala and Kafka. | 2024-12-18 | |
EMR 5.17.4 & EMR 3.51.4 release | This release upgrades multiple services and includes fixes and optimizations for previous versions. | 2024-12-18 |
November
Feature | Description | Release date | Related documentation |
Support for cluster-level release protection | When creating a cluster, you can enable release protection to prevent accidental resource loss. | 2024-11-28 | |
Support for upgrading pay-as-you-go node groups | You can upgrade a pay-as-you-go node group to handle high loads during business demand spikes. | 2024-11-28 | |
Support for elastic capacity assurance for pay-as-you-go node groups | You can reserve ECS resources in advance by associating them with a private pool. When you create a pay-as-you-go node group, you can link it to this pool to ensure resource availability. | 2024-11-28 | |
Support for system disk encryption | When you create a cluster, you can encrypt the system disk by using a key from KMS. | 2024-11-28 | |
Bootstrap action optimization | Bootstrap actions can now run after component installation but before service startup. | 2024-11-28 |
October
Feature | Description | Release date | Related documentation |
Support for managed auto scaling | When managed auto scaling is enabled, the system continuously monitors the cluster's YARN load. You specify the minimum and maximum number of task nodes, and the system adjusts the node count based on the load to maximize resource utilization. | 2024-10-22 |
August
Feature | Description | Release date | Related documentation |
New monitoring and diagnostics feature | This new monitoring and diagnostics feature for EMR on ECS is an intelligent O&M tool powered by large models. It leverages the expertise of the Alibaba Cloud EMR team and advanced diagnostics to enhance platform observability. The feature provides real-time cluster health diagnostics, performs root cause analysis for anomalies, and suggests solutions to reduce O&M costs. It also generates daily cluster reports with optimization insights to improve efficiency. | 2024-08-20 | |
Cluster cloning optimization | Cluster cloning now restores modified service configurations, added node groups, and configured auto scaling rules. This allows you to quickly replicate a cluster's configuration. | 2024-08-20 | |
Increased limit for additional security groups per node group | Each node group can now be associated with up to four additional security groups for more flexible network access control of its ECS instances. | 2024-08-20 |
June
Feature | Description | Release date | Related documentation |
Support for auto-renewal during scale-out | If auto-renewal is enabled for a cluster, newly added nodes also have it enabled by default. You can manage the auto-renewal status and duration at any time. | 2024-06-19 | |
Support for converting pay-as-you-go node groups to subscription | You can convert pay-as-you-go core, task, and gateway node groups in a subscription cluster to the subscription billing method for more flexible billing. | 2024-06-19 | |
Support for Master-Extend node groups for custom deployments | You can create Master-Extend node groups to deploy components for services like Spark, Hive, and Kyuubi, with automatic configuration synchronization. This reduces the load on the master node group. | 2024-06-19 |
March
Feature | Description | Release date | Related documentation |
Support for managing OSS-HDFS buckets in the EMR console | You can now create an OSS-HDFS bucket in the EMR console while creating a cluster. You can then view the bucket's storage overview and file list from the cluster's service page without switching to the OSS console. This simplified workflow helps prevent operational errors that could cause HDFS service outages. | 2024-03-14 | |
Support for gateway node groups | You can add gateway node groups to a cluster to act as task submission machines and offload the master node. This feature enables one-click creation of task submission machines with automatic configuration synchronization, simplifying deployment. | 2024-03-14 | |
Health check item management | E-MapReduce automatically runs health checks on cluster nodes and services to detect potential issues early. You can view check details for nodes and service components, and edit the health check items. | 2024-03-14 | |
Expanded health checks for service components | This update adds more health check items for YARN, HDFS, Hive, Kafka, and ZooKeeper to improve health monitoring accuracy. | 2024-03-14 |
2023
October
Feature | Description | Release date | References |
Auto scaling rule recommendation optimization | The recommendation algorithm for auto scaling rules has been optimized. It now provides more precise and effective suggestions based on your cluster's resource utilization, further enhancing resource elasticity and cost-effectiveness. | 2023-10-24 | |
New alert management feature | The cluster management page includes an alert management feature, which is based on CloudMonitor. In the EMR console, you can create and view alert rules for your cluster. If a resource's monitoring metric reaches the alert threshold, CloudMonitor automatically sends an alert notification, enabling you to promptly identify and address cluster exceptions. | 2023-10-24 | |
New node health status feature | The node health status indicates whether a node is running correctly. View the node health status on the node management page to promptly identify abnormal nodes. | 2023-10-24 | |
Support for configuring ESSD performance levels | When you create a cluster or add a node group, you can set different performance levels (PLs) for ESSDs to meet various cluster performance requirements. | 2023-10-24 |
August
Feature | Description | Release date | References |
New cluster template feature | A cluster template saves the configurations of an EMR instance, allowing you to create EMR clusters quickly. | 2023-08-29 | |
New cluster resource overview | The auto scaling module now includes a cluster resource overview. This overview analyzes resource utilization and recommends auto scaling rules, allowing you to enable auto scaling and improve your cluster's resource elasticity. | 2023-08-29 | |
Enhanced visibility for configuration items | When a configuration item is modified at the node group or node level, the cluster-level view now displays multi-level details, including the specific node group or node name and its configuration. | 2023-08-29 |
July
Feature | Description | Release date | References |
New auto scaling management module | To simplify cluster elasticity management, EMR has added a dedicated auto scaling management module. In this module, you can manage scaling rules and view the usage and costs of elastic resources. This allows you to evaluate cost savings from auto scaling and optimize cluster resource utilization. | 2023-07-12 | |
Automatic compensation optimization | The automatic compensation feature replaces abnormal nodes in a cluster. This optimization adds informational messages and event notifications, allowing you to track the automatic compensation process. Note
Starting from 18:00 (UTC+8) on July 10, 2023, the Automatic Compensation switch is enabled by default when you create a pay-as-you-go task node group. | 2023-07-12 | |
Service configuration optimization | New 'Pending Deployment' and 'To Take Effect' indicators have been added to better guide users on the actions required after a configuration is modified, preventing situations where the changes do not take effect. | 2023-07-12 | |
Support for stateless clusters | EMR now supports a default data lakehouse architecture that does not depend on HDFS. If you do not use services that rely on a core node group, you can remove it to build a fully stateless cluster and reduce O&M costs. | 2023-07-12 | |
Support for YARN partition and queue association | You can manage YARN partition-to-queue associations and capacity allocations directly in the console, replacing tedious manual configuration. | 2023-07-12 | |
Per-second billing for some pay-as-you-go resources | More granular billing provides more accurate charges and reduces your usage costs. | 2023-07-12 |
June
Feature | Description | Release date | References |
Version updates |
| 2023-06-01 | |
New Paimon component | Apache Paimon is a unified streaming and batch data lake storage format that supports high-throughput writes and low-latency queries. | 2023-06-01 | |
New Presto component | Presto (also known as PrestoDB) is a flexible and scalable distributed SQL query engine. | 2023-06-07 |
April
Feature | Description | Release date | References |
Version updates |
| 2023-04-03 | |
New data lakehouse capabilities | EMR allows the Spark and Trino compute engines to directly access Hologres and MaxCompute tables. | 2023-04-03 | New data lakehouse capabilities: EMR supports Hologres and MaxCompute data sources |
Spark integration with Hologres | You can use Spark to read data from and write data to Hologres tables. | 2023-04-03 | |
Upgrade node configurations | You can upgrade the ECS instance type for a node group. | 2023-04-03 | |
Manage YARN partitions in the console | You can manage YARN partitions and map multiple node groups to them in batches through the console's visual interface. | 2023-04-13 |
March
Feature | Description | Release date | References |
New Flink Table Store component | Flink Table Store is a unified streaming and batch data lake storage format that supports high-throughput writes and low-latency queries. | 2023-03-03 | - |
Export and import service configurations | You can export service configurations in XML or JSON format to back up, migrate, and restore them. | 2023-03-02 |
February
Feature | Description | Release date | References |
Version updates |
| 2023-02-28 |
2022
December
Feature | Description | Release date | Related documentation |
Version update |
| 2022-12-01 | |
YARN node labels | YARN node labels let you manage YARN NodeManager nodes in partitions. | 2022-12-14 |
November
Feature | Description | Release date | Related documentation |
Version update |
| 2022-11-08 | |
Log management | You can query the logs of open source components directly in the E-MapReduce console. | 2022-11-29 |
October
Feature | Description | Release date | Related documentation |
Version update |
| 2022-10-14 | |
HBase Shell | You can use HBase Shell to connect to HBase. | 2022-10-21 | |
Data serving cluster | A data serving cluster is an E-MapReduce cluster type built on Apache HBase. | 2022-10-28 |
September
Feature | Description | Release date | Related documentation |
Auto recovery | When auto recovery is enabled, E-MapReduce (EMR) monitors your cluster's ECS instances. If an instance becomes faulty, the feature attempts to replace it with a new one. | 2022-09-07 | |
Cluster cloning | Create new clusters quickly by cloning existing ones. | 2022-09-09 |
August
Feature | Description | Release date | Related documentation |
Version update |
| 2022-08-05 | |
Deployment set | A deployment set is an Alibaba Cloud Elastic Compute Service (ECS) feature that controls the distribution of ECS instances to improve disaster recovery and availability. | 2022-08-05 | |
Custom gateway deployment with EMR-CLI | E-MapReduce provides EMR-CLI, a tool to deploy a gateway environment on Alibaba Cloud ECS instances. You can create ECS instances as needed and use EMR-CLI to quickly deploy the environment. | 2022-08-05 |
July
Feature | Description | Release date | Related documentation |
EMR Doctor | EMR Doctor is an intelligent diagnostics system for open source big data clusters in E-MapReduce. | 2022-07-25 |
June
Feature | Description | Release date | Related documentation |
Data lake cluster | Data lake clusters are a flexible, reliable, and efficient type of big data computing cluster available in E-MapReduce. | 2022-06-01 | |
Spark cluster association with RSS | EMR Remote Shuffle Service (RSS) is an E-MapReduce component that improves the stability and performance of Spark's native shuffle process. You can associate Spark clusters on EMR on ACK with an RSS cluster. | 2022-06-09 |
May
Feature | Description | Release date | Related documentation |
Memory management | Manage and monitor memory for the StarRocks backend (BE), including configuration and usage details. | 2022-05-10 |