E-MapReduce:Console release notes

Category

Description

Instance management

• New features:

  • You can now create compute groups across availability zones.

  • AI Center now supports the Qwen 3.5-plus and Qwen 3.6-plus models.

  • AI Center now displays token usage statistics.

StarRocks Manager

• Enhancements:

  • Fine-grained access control now lets you manage user permissions for downloading query results.

Category

Description

Instance management

• New features:

  • Compaction Service now supports cross-zone deployment.

  • You can now set an upper limit for the mem-limit parameter.

  • Monitoring and alerting:

    • You can now deliver raw monitoring metric data to SLS for custom monitoring scenarios.

    • You can now configure alerts for a materialized view that becomes invalid or inactive.

• Enhancements:

  • Alert rule configurations now include the instance name to help identify alert sources.

  • Improved the auto scaling logic to prevent frequent scaling operations from causing scaling to stall.

Category

Description

Instance management

• New features:

  • Introduces the Compaction Service (for StarRocks 3.5 only), a flexible and isolated service for compaction. This feature is in beta testing.

  • Monitoring and alerting:

    • You can now deliver raw monitoring metric data to SLS for custom monitoring scenarios.

    • You can now configure alerts for a materialized view that becomes invalid or inactive.

• Enhancements:

  • The content for business insights (including query insights, import insights, and cache insights) is restructured and optimized. For details, see the corresponding documentation.

  • The All Operations menu for instances is now optimized.

StarRocks Manager

• Enhancements:

  • The SQL Editor now defaults to the current user's specified compute group when opened.

  • You can now view the number of rows scanned in an SQL task.

  • The Profile generation logic has been optimized to fix issues that prevented Profile generation in some scenarios.

  • Viewing size and row count metrics for external tables is no longer supported. This change prevents aggressive scanning to protect the stability of external metadata services, such as HMS.

Category

Description

instance management

• New features:

  • You can now switch from subscription to Pay-As-You-Go.

  • You can now add tags to compute groups to track their costs in your bills.

  • You can now use All Operations to act on all instances at once.

  • You can now associate an instance with DLF during instance creation.

  • AI Function is now available in beta.

  • You can now manage permissions for compute groups in Ranger.

• Enhancements:

  • The system now recommends upgrades for critical defects in the software stack, via highlighted alerts and event notifications.

  • The Diagnosis Daily Report under health diagnosis now includes new checks that verify whether index persistence for primary key tables is enabled, whether the index type is cloud_native, and whether the bucket size is less than 5 GB.

  • Added an inventory pre-check for scale-out and scale-up operations.

  • Real-time Diagnosis now supports filtering by compute group.

  • You can now modify a scaling policy while a scaling event is in progress.

• Resolved issues:

  • Fixed an issue that could cause the system disk to become full.

  • Removed an invalid chart from the compaction group view in Monitoring for storage-compute-separated instances.

StarRocks Manager

• New features:

  • You can now associate users with a RAM role to use DLF's RAM role-based permissions.

• Enhancements:

  • In SQL Tasks > Running Tasks, you can now terminate multiple tasks at once and sort tasks by field.

  • The Import Tasks page now supports filtering and includes a task duration field.

Feature category

Description

Instance management

• New features:

  • The Event Center now integrates with Cloud Monitor (CMS), letting you view all StarRocks cluster events from a single location.

  • You can now use alert templates to apply basic alert configurations with a single click.

  • New Cloud Monitor metrics are now available for alert configurations, including query, transaction, database lock, node status, FE memory, and MV.

  • The new health diagnosis daily report automatically assesses your instances' health. It provides evaluations, repair suggestions, and optimization best practices for resource configuration, query tasks, import tasks, compaction, table analysis, and cache analysis.

  • You can now collect and export cost statistics by the instance and compute group dimensions.

  • The Terraform Provider now supports elastic scaling for StarRocks clusters.

  • Partial graceful shutdown is now supported when scaling in cloud-native clusters. By default, the system waits 30 seconds after adding a node to the denylist before scaling in.

  • When scaling in an all-in-one cluster, you can now select and decommission multiple BE/CN nodes in parallel. This operation carries a high risk.

  • BE nodes now expose port 9060 to enable public network access from Spark to StarRocks instances.

• Feature optimizations:

  • You can now modify the emr_internal_redirect parameter online in the console. This allows Stream Load to use V2 and avoid SLB streaming.

• Bug fixes:

  • Fixed an issue where the system disk would become full.

  • Removed invalid charts from the compaction group in the monitoring section for cloud-native instances.

StarRocks Manager

• Feature optimization:

  • Improved task status reporting to correctly reflect successful completion and prevent false timeout errors.

• Bug fix:

  • Fixed an issue where successfully executed queries were sometimes reported as failed. For example, adding LIMIT 100 to the query SQL would return results and show a normal task status.

Documentation

• Published the Critical bug notifications document.

• Published the Version lifecycle policy document.

• Added three parameters to the Usage limits document to restrict modifications during the table design stage.

Feature category

Description

instance management

• New features:

  • The Event Center feature is now generally available. It provides notifications for underlying events from services such as ECS and OSS.

  • Elastic ephemeral disks are now supported for compute-storage separated instances. This feature is in beta and available only in specific regions and zones.

  • You can now scale in cloud disks for compute groups in compute-storage separated instances.

  • The multi-zone disaster recovery feature is now generally available for compute-storage separated instances.

  • The new Health Diagnosis feature is now available. It includes the existing Real-time Diagnosis module and adds the new Diagnostic Daily feature, which is currently in beta. Diagnostic Daily provides an intuitive health score, along with assessment results, recommended actions, and best practices for various dimensions.

  • You can now enable SLB and a standalone agent when you create an instance.

  • The EMR StarRocks API now supports querying a list of instances.

• Enhancements:

  • Auto Scaling now supports multiple scaling policies.

  • The Health Report feature has been renamed to Business Insight. This feature provides multi-dimensional reports to support data governance and business administration.

  • A new report, "Tables Not Accessed in the Last 90 Days," has been added to Business Insight.

  • Log query now lets you select specific log files for delivery to reduce SLS costs.

  • Optimized the default JAVA_OPTS parameters.

  • Optimized the restart process.

  • Optimized the data sources for CPU and memory consumption in Real-time Diagnosis.

  • Improved the user experience on the Real-time Diagnosis page.

  • Optimized several monitoring dashboards and added new monitoring metrics.

  • Improved the user experience for selecting vSwitches and zones when you create an instance.

  • Alert Monitoring now supports quick time range selection.

• Bug fixes:

  • Fixed an issue where FE disks could become full.

  • Fixed an issue that could block the instance deletion process.

StarRocks Manager

• New features:

  • Added 1 TB TPC-H and TPC-DS test templates to SQL Editor.

• Enhancements:

  • You can now set user properties when you create a user.

  • You can now associate existing StarRocks users with RAM users to enable permission integration with DLF.

  • Audit logs now support the SQL fingerprint feature.

• Bug fixes:

  • Fixed an issue where the browser would occasionally stutter on the SQL task list page.

Feature category

Description

Instance management

• Added the Compute enhanced CU type. With a 1:2 vCPU-to-memory ratio, it is ideal for high-CPU workloads. This feature is available to whitelisted users only.

• You can now create instances using Terraform.

Feature category

Description

Instance management

• New features:

  • You can now use cloud disk encryption and Secure Sockets Layer (SSL) encryption on integrated storage and compute instances to enhance security.

  • FE nodes now support the memory-enhanced instance family, which is ideal for scenarios with high FE memory usage.

  • Integrated storage and compute instances now support fast restart.

  • You can now configure a custom domain name for a StarRocks instance to access its custom endpoint.

  • For subscription instances, you can now convert pay-as-you-go compute groups to the subscription model.

  • You can now apply coupons for discounts on instance purchases.

• Feature optimizations:

  • Improved the user experience on the instance purchase page.

  • Improved the display of configuration items for each software stack version in the parameter settings.

  • You can now view new alert rules configured in CloudMonitor on the Monitoring and Alerting > Alert Settings tab in the console.

  • You can now manage StarRocks with Apache Ranger.

  • You can now add descriptions to compute groups.

  • The Instance Details page now includes a link to the kernel release version.

  • Additional monitoring metrics:

    • View the memory usage of FE nodes.

    • View historical trends for the average CPU, memory, and disk usage of BE nodes.

StarRocks Manager

• New feature:

  SQL Editor query results now include corresponding data types and support null values.

• Feature optimizations:

  • Provided documentation on executing SQL statements in the SQL Editor. For statements that contain a semicolon (;), you must add $$ as the delimiter.

  • Improved the display of the runtime end time for asynchronous materialized view tasks on integrated storage and compute instances.

  • Refined the content on the Manager Overview page for better clarity.

• Bug fixes:

  Optimized metadata management: Fixed an issue where Paimon Catalog sometimes failed to load or display data correctly.

Feature category

Description

Instance management

• New features:

  • You can now query detailed instance logs to troubleshoot task exceptions.

  • You can now use gateway management to group and isolate FE nodes and restrict the leader node to write operations. You can also enable SLB load balancing to enhance instance stability.

  • You can now scale in FE nodes when your instance has more than three nodes.

  • You can now get the migration token from the Instance Details page.

  • The Japan (Tokyo) region is now available.

• Enhancements:

  • Enhanced the restart notification for instance upgrades.

  • Refined the user experience on the real-time diagnostics page.

  • The latest minor version is now displayed when you create an instance.

  • Added configurations for Kerberos and Ranger.

  • Enhanced the user experience of monitoring charts.

  • Enhanced the console overview page.

  • You can now apply configuration changes without an immediate restart, and combine the restart with other operations later to reduce service downtime.

StarRocks Manager

• You can now search within the Catalogs area on the Metadata Management page.

• Enhanced the performance of SQL query history without affecting the current instance.

• The user page now includes a search box to help you find users quickly.

• Enhanced the user experience of the import task filter.

• Refined the content display on the query details page.

• Enhanced the user experience of SQL Editor.

Feature category

Description

Instance Operations

• New Features:

  • Supports new CU types, including memory-enhanced (RCU) and network-enhanced (NCU), to meet diverse performance requirements.

  • Operations requiring a restart, such as applying configuration changes and instance scaling, now use fast restart.

  • Supports password reset for the admin user.

  • Shared-data instances now support one CN node.

• Enhancements:

  • Improved real-time diagnostics and health reports.

  • Optimized the version upgrade process.

  • Optimized monitoring and alerting metrics:

    • Added support for new monitoring metrics, including FE disk metrics and compute node IOUtil.

    • Storage-related metrics are now on a dedicated storage tab.

StarRocks Manager

• New Features:

  • Metadata management: Supports visually creating tables, databases, and materialized views, and viewing materialized view information.

  • Security Center: Supports visual management of users, roles, and permissions using Role-Based Access Control (RBAC).

  • SQL Editor: Supports configuration of the LIMIT clause.

• Enhancements:

  • Enhanced the user experience of SQL Editor.

  • Optimized the display of SQL Profile.

  • Improved sorting and search for tables in metadata management.

Feature category

Description

Control platform features

• Shared-nothing instances now support backup and recovery (in public preview).

• You can now change the time zone.

• Enhancements to the health report feature:

  • A cache analysis report is now available.

  • SQL query analysis now includes Query Latency P99 and DB Lock metrics.

• The SQL profile now includes "Top 10 by Time" and "Top 10 by Memory" sections.

• Slow query logs now include fields for CPU and memory usage.

Feature category

Update description

control platform capabilities

• You can now manage your instances using Alibaba Cloud resource groups and tags.

• You can now view materialized views and their task information in StarRocks Manager.

• You can now run ready-to-use TPC-DS and TPC-H test scripts directly in the SQL Editor of StarRocks Manager.

• You can now view the records and details of INSERT INTO tasks directly under Import Tasks or SQL Tasks in StarRocks Manager.

• Enhancements:

  • The metrics in the Profile have been enhanced.

  • The low disk space notification has been improved.

Feature category

Description

Control platform capabilities

• Instances now support auto scaling based on a schedule. This feature is currently limited to pay-as-you-go instances.

• Added new monitoring metrics and optimized monitoring reports.

StarRocks Manager enhancements

• SQL Profile now supports diagnostic suggestion cards to analyze and optimize SQL queries.

• Queries that fail due to syntax errors are now displayed in the All Queries view.

Experience improvements

• Improved user prompts in the purchase workflow.

• Improved the QueryId link and its text on the health report page.

• Improved the text on the SQL Profile interface.

• Improved the performance of the SQL Editor.

Feature category

Description

Control platform capabilities

• Adds support for three new instance types: shared-nothing, Data Lake Analytics Edition, and shared-data.

• Enhances diagnostic analysis by adding support for SQL query analysis, table analysis, and import task analysis to help you quickly identify problematic tasks or data.

• Improves the user experience for the instance configuration page.

Feature category

Description

Compute and storage separation

• Metadata management now displays databases, tables, partitions, and cache.

• Query Profile now includes cache hit analysis metrics.

Data lake analytics

• Metadata management now displays information about external tables from an external catalog.

• You can now configure password-free access for OSS.

Control platform enhancements

• You can now change cloud disk configurations, including upgrading the performance level.

• You can now upgrade a cloud disk configuration by increasing the number of blocks.

• You can now adjust the number of cloud disk blocks when creating an instance.

• Optimized the default configuration items.

• The monitoring feature now lets you query hotspot monitoring metrics.

• Operation tasks now display operation details.

• A confirmation prompt now appears when enabling an audit log.

StarRocks Manager enhancements

• Improved the connection and sign-in experience for StarRocks Manager.

• StarRocks Manager now supports light and dark themes.

• The user management feature now includes a Description field.

Regions

The China (Zhangjiakou) region is now available.