Classmethod Data Analytics Newsletter (AWS Data Analytics Edition) – September 2026 Issue
This page has been translated by machine translation. View original
This is Ishikawa from the Cloud Business Headquarters. Here is the AWS data analytics update information for August 2026. This month, notable highlights included the general availability of AWS Glue 6.0 with a 30% price reduction and Apache Iceberg v3 support, the introduction of streaming tables enabling direct delivery from Amazon Kinesis Data Streams to Amazon S3 Tables, and Amazon Redshift's long-term system table retention and Apache Iceberg v3 support. It was a month where, centered around Iceberg v3 and S3 Tables, each service aligned toward reducing the need for custom pipeline implementations through managed integrations. There are other updates to introduce as well!
Amazon Redshift
New Features & Updates
2026/08/20 - Amazon Redshift introduces long-term system table retention with Amazon S3 Tables integration
Through native integration with Amazon S3 Tables, system table data can now be retained long-term beyond the conventional 7-day period. Once enabled, system table data is automatically written to S3 Tables in Apache Iceberg format, with AWS managing partitioning, compaction, and retention periods.
Previously, extending the retention period required building and maintaining custom ETL pipelines to copy system tables. When operating multiple data warehouses, there was also the complexity of aggregating system tables from each warehouse using data sharing. This new feature eliminates the need for custom ETL.
2026/08/20 - Amazon Redshift now supports concurrency scaling of streaming ingestion workloads from Amazon Kinesis data streams
Starting with Amazon Redshift patch P203 and later, concurrency scaling is now applied to updates of streaming materialized views connected to Amazon Kinesis Data Streams.
Redshift streaming ingestion is a feature that ingests data from KDS with low latency, storing data in streaming materialized views. With concurrency scaling enabled, streaming workloads automatically scale, freeing up the main cluster or workgroup for other higher-priority workloads. Available in all regions where Redshift is available.
2026/08/27 - Amazon Redshift integrates with Agent Toolkit for AWS for AI-assisted data warehouse management
Amazon Redshift has integrated with Agent Toolkit for AWS, enabling AI agents such as Claude Code, Kiro, and Cursor to build, query, troubleshoot, and migrate Redshift data warehouses and data lakes.
It is provided as a combination of an AWS MCP server that is authenticated and executes AWS APIs, and Redshift skills, which is a curated collection of procedures and references for Redshift. The skills include SQL syntax references, metadata exploration, data load patterns, materialized view best practices, explanations of extensions such as Qualify, Pivot, and Super, and step-by-step migration procedures.
2026/08/27 - Amazon Redshift streaming can now ingest 10MiB records from Amazon Kinesis Data Streams
Amazon Redshift streaming ingestion now supports a maximum record size of 10MiB from Amazon Kinesis Data Streams. This is a tenfold increase from the previous 1MiB limit, aligning with the expanded maximum record size on the KDS side.
Since large payloads can be ingested into Redshift as-is without record splitting, ingestion pipelines can be simplified, and high-volume, large-record workloads can be accommodated. Available in all commercial regions where Amazon Redshift is available.
2026/08/31 - Amazon Redshift now supports AWS IAM Identity Center authentication with enhanced VPC routing
AWS IAM Identity Center authentication is now available for provisioned clusters and serverless workgroups configured with Enhanced VPC Routing (EVR). Single sign-on using corporate credentials can be used while keeping communications within the AWS network via the VPC.
Redshift validates and exchanges IAM Identity Center tokens from within the VPC via AWS PrivateLink interface VPC endpoints, so authentication and authorization also follow the same managed network path as other Redshift communications. This is useful for environments where communication to the public internet is not permitted due to data residency or network isolation requirements.
2026/08/31 - Amazon Redshift now supports Apache Iceberg v3 tables
Amazon Redshift now supports reading and writing Apache Iceberg v3 tables on the data lake. This release adds support for default column values, row lineage, and deletion vectors.
Default column values allow you to define initial values applied when no value is specified, making schema evolution easier when adding columns. Row lineage exposes pseudo-columns that track each row's identifier and last update sequence number, enabling you to build incremental pipelines and CDC workflows that process only changed rows. Deletion vectors replace v2 positional delete files with compressed bitmaps, speeding up reads and writes for workloads with many updates and deletes.
API Changes
2026/08/19 - Amazon Redshift - 20 updated methods
Support has been added for Enhanced System Table Retention, which stores system table data directly in Amazon S3 Tables within the customer account rather than in Redshift Managed Storage.
Amazon Redshift Serverless
New Features & Updates
API Changes
2026/08/19 - Redshift Serverless - 7 updated methods
Support has been added for Amazon Redshift Enhanced System Table Retention, which stores system table data directly in Amazon S3 Tables within the customer account rather than in Redshift Managed Storage.
AWS Glue
New Features & Updates
2026/08/05 - AWS Glue Data Catalog now supports metadata exports to S3 Tables (Preview)
Two features have been added to the AWS Glue Data Catalog currently in preview: export of catalog metadata to Amazon S3 Tables, and expanded semantic search preview for catalogs encrypted with AWS KMS customer managed keys.
When export is enabled, technical and business metadata such as glossary terms, custom metadata attachments, and asset descriptions are written to tables in a managed aws-catalog S3 table bucket. Since data is stored in Apache Iceberg format, you can execute queries, audits, and time-travel queries using standard SQL from Amazon Athena, Amazon QuickSight, and Iceberg-compatible third-party tools.
2026/08/05 - AWS Glue Data Quality makes ETL anomaly detection free and improves anomaly predictions
A new observation mode has been added to AWS Glue Data Quality anomaly detection to reduce false positives, and the charge for anomaly detection in ETL jobs has been eliminated. By using a constant baseline rather than a linear trend, over-extrapolation of trends can be avoided, enabling more accurate alerts.
It is well-suited for cases where data arrival intervals are irregular, exploratory data analysis, datasets with flat or random patterns, workloads without predictable trends, and execution in interactive environments such as notebooks. Available in all AWS commercial regions and AWS GovCloud.
2026/08/06 - AWS Glue Schema Registry is now available in ten more AWS regions
AWS Glue Schema Registry is now available in 10 regions: Asia Pacific (New Zealand, Thailand, Hyderabad, Osaka, Malaysia, Melbourne, Taipei), Mexico (Central), Israel (Tel Aviv), and Canada West (Calgary).
Schema Registry is a serverless and free feature of AWS Glue that registers Apache Avro, JSON, and Protobuf schemas to validate and control the evolution of streaming data structures. Through Apache-licensed serializers/deserializers, it integrates with Apache Kafka/Amazon MSK, Amazon Kinesis Data Streams, Apache Flink/Amazon Managed Service for Apache Flink, and C#/Java applications for AWS Lambda.
2026/08/11 - AWS Glue adds one-click access to SageMaker Unified Studio from the AWS console
One-click access to Amazon SageMaker Unified Studio from the AWS Glue console is now available. From a state where you are browsing the catalog in the Glue console, you can open SageMaker Unified Studio with the same IAM role and move to query execution, data quality checks, and building data pipelines.
With this addition, SageMaker Unified Studio can now be opened with one click from the S3 Tables, Athena, EMR, Redshift, and Glue consoles. If SageMaker Unified Studio is not yet set up, the necessary IAM policies can be created and configured from the inline permissions panel without navigating to the IAM console.
2026/08/21 - AWS Glue 6.0 delivers 30% price reduction and Iceberg v3 support
AWS Glue 6.0 is now generally available. In addition to a 30% price reduction, it introduces full support for Apache Iceberg v3, support for new versions of Apache Hudi and Delta Lake, and features that improve development productivity. The runtime has been updated to Apache Spark 4.1, Python 3.13, and Scala 2.13.
For Iceberg v3 support, the VARIANT type with automatic shredding for faster reading of semi-structured data, deletion vectors for faster row-level updates, geometry/geography types for spatial processing, and flexible schema evolution with the UNKNOWN type and default column values have been added. Additionally, Spark Declarative Pipelines that eliminate the need for boilerplate orchestration code, Real-Time Mode streaming with sub-second latency, and Arrow-native Python UDFs have also been introduced.
API Changes
2026/08/05 - AWS Glue - 2 new methods
PutDataCatalogExportConfiguration has been added for exporting Glue Data Catalog metadata to system tables on Amazon S3 Tables.
2026/08/14 - AWS Glue - 3 updated methods
It is now possible to associate glossary terms with form items that have repeated items, such as table columns.
Amazon Quick
New Features & Updates
2026/08/06 - Amazon Quick supports multi-dataset analytical capabiity
Amazon Quick now supports multi-dataset topics. You can define relationships between multiple datasets within a single topic and use that model to create dashboards and answer questions in natural language.
Previously, visualizations spanning datasets required pre-joining, which involved implementing JOINs in data preparation, consuming additional SPICE capacity, and recreating datasets for each use case or model change. With multi-dataset topics, the topic becomes a reusable relational data model, and once relationships are defined, Quick performs the joins at runtime.
2026/08/14 - Amazon Quick now supports approval policies for sharing
Approval policies have been added to Amazon Quick, enabling administrators to govern asset sharing within the organization. It is now possible to operate in a way that does not grant access until designated approvers review and approve sharing requests, keeping sensitive asset sharing intentional and auditable.
Policies can be scoped by asset type, such as knowledge bases, spaces, and custom chat agents. Approvers can review the assets themselves before approving or rejecting, and all workflow events are recorded in AWS CloudTrail. Available with Professional and Enterprise plans.
2026/08/14 - Amazon Quick now supports customizable no data messages for visuals
Authors can now replace the default "No data to display" message shown when no data is returned for a visual with any content of their choice. A title, subtitle, and hyperlink can be configured, communicating to readers why data is not displayed and what they should do next.
All text fields support parameters, so filter content selected by the reader can be dynamically reflected in the message. Hyperlinks support http://, https://, and mailto:, and the title and subtitle can each be toggled on or off independently. Available in all regions where Amazon Quick is offered.
2026/08/14 - Amazon Quick now supports data loss prevention with Microsoft Purview
Amazon Quick has integrated with Microsoft Purview, enabling data loss prevention (DLP) policies to be applied across the Quick environment. Existing Purview sensitivity labels can be used to automatically control the handling of files across features such as chat, spaces, and knowledge bases.
Administrators can configure enforcement actions such as block, warn, or allow for each sensitivity label. For example, files labeled "Highly Confidential" can be blocked from being uploaded to shared spaces, while "Internal" can be allowed with a warning. Existing Purview governance policies can be extended to Quick without any additional tools.
2026/08/14 - Amazon Quick now supports per-user resource limits
Administrators can now set limits on index storage and agent time on a per-user basis. By creating limit profiles that restrict per-user consumption, unexpected overcharge fees can be prevented while subscription entitlements can be used efficiently across the organization.
Limit profiles can be created and assigned at the user, role, and account levels, with an appropriate capacity applied through a priority hierarchy. Users who reach their limit are blocked from new consumption, but existing content is retained. Available with Professional and Enterprise plans.
2026/08/17 - Amazon Quick Microsoft 365 extensions are now generally available
Microsoft 365 extensions for Excel, PowerPoint, Word, and Outlook are now generally available. Quick can directly perform tasks within the user's M365 environment, handling tasks such as document redlining, financial model creation, presentation material creation, and Outlook inbox management.
The Excel extension assists with complex aggregations, creating pivot tables and charts, and data import and cleansing. The PowerPoint extension creates materials from Quick data in line with organizational templates, and the Word extension supports bulk editing with tracked changes and participating as a reviewer with comments.
2026/08/19 - Amazon Quick adds deny by default for custom permissions
"Deny by default" has been added to Amazon Quick's custom permissions. This is a governance setting that automatically restricts new AI features before they reach users, eliminating the need for administrators to reactively respond after a feature release as was previously required.
By restricting AI feature categories in custom permission profiles and assigning them to users, roles, and accounts, AI features released thereafter are denied by default. Restricting a category also restricts existing features within that category. The operation becomes one of explicitly permitting features as they are ready.
2026/08/31 - AWS Agent Registry agents and MCP servers now available in Amazon Quick
Amazon Quick has integrated with AWS Agent Registry, making it possible to search and browse agents and MCP servers registered in the organization's Agent Registry from within Quick and enable them in just a few clicks. Connection information is automatically pulled from the registry.
Enabled resources can be shared within a team and used across chat, agents, apps, flows, and deep research. It serves as a bridge between technical teams building agents with Amazon Bedrock AgentCore and business users using Quick. Available in all regions where both Amazon Quick and Amazon Bedrock AgentCore are available.
API Changes
2026/08/12 - Amazon QuickSight - 16 new 4 updated methods
APIs have been added for DLP using Microsoft Purview (configuration management for label enforcement across spaces, chat, and knowledge bases), approval workflows (CRUD for sharing policies for agents, knowledge bases, and spaces), and limit management (limit profiles for index storage and agent time per user).
2026/08/31 - Amazon QuickSight - 6 new methods
ListApps, SearchApps, DescribeApp, DescribeAppPermissions, UpdateAppPermissions, and DeleteApp have been added for Amazon QuickSight app management.
Amazon SageMaker Unified Studio
New Features & Updates
2026/08/18 - Amazon SageMaker Unified Studio now supports data profiling and anomaly detection
Data profiling and anomaly detection powered by AWS Glue Data Quality are now supported. Statistical profiles can be generated to understand the shape and missing state of data, and how those statistics have changed over time can be tracked. Anomaly detection detects deviations from historical patterns without pre-defining thresholds or custom rules.
It supports both stored data on catalog tables and data being processed within Visual ETL jobs. A dedicated Data profile tab has been added to catalog tables, calculating per-dataset and per-column statistics on demand and on a schedule. As profile history accumulates, a baseline of expected behavior is built up.
2026/08/19 - Amazon SageMaker notebooks now support trusted identity propagation
SageMaker notebooks now support Trusted Identity Propagation (TIP) for Amazon Athena, Amazon Redshift, and Amazon EMR Serverless. When connected to TIP-enabled compute in a TIP-enabled project, the IAM Identity Center identity of each notebook user is propagated to AWS Lake Formation, making only the tables, columns, and rows for which they have permissions visible.
There is no longer a need to share a single execution role with broad permissions, and per-user data boundaries based on the query executor and an audit trail of access principals via CloudTrail are obtained. Identities are automatically propagated through existing compute connections, requiring no additional logins, tokens, or role management. Available in all regions where Amazon SageMaker Unified Studio is available.
Amazon DataZone
New Features & Updates
API Changes
2026/08/11 - Amazon DataZone - 15 updated methods
GetSubscriptionGrant now returns materialized asset scope names for mapping AWS Lake Formation data cell filters and Amazon Redshift views corresponding to subscription grants.
2026/08/27 - Amazon DataZone - 2 updated methods
cascadeDelete has been added to DeleteDomain. When specified, it recursively deletes projects, environments, subscriptions, and underlying AWS resources before deleting the domain. Progress can be checked with deleteProgress in GetDomain, and failure reasons per resource with failureReasons.
Amazon EMR
New Features & Updates
2026/08/04 - Run interactive workloads on Amazon EMR on EC2 with Spark Connect
Amazon EMR on EC2 now supports interactive Apache Spark sessions via Spark Connect. Sessions can be run on a dedicated EMR on EC2 cluster from managed notebooks in Amazon SageMaker Unified Studio or familiar IDEs such as Jupyter and Visual Studio Code.
Since a persistent Spark context that persists across cells and scripts can be maintained, it is well-suited for ad-hoc exploration combining local Python code execution with remote Spark processing, step-by-step debugging, and iterative development of PySpark jobs before production deployment. Running and completed sessions can be monitored and debugged from the EMR console.
2026/08/28 - Amazon EMR on EKS now supports job run concurrency controls
Admission control has been added to Amazon EMR on EKS to control the number of job executions per virtual cluster. The maximum number of concurrently running jobs maxConcurrentJobRuns and the maximum number of jobs that can wait in the queue maxInQueueJobRuns can optionally be configured, and backpressure signals can be received from the StartJobRun API.
This prevents the EKS cluster shared in a multi-tenant environment from being overloaded and prevents scheduling failures caused by noisy neighbors, allowing important workloads to run in a predictable manner even under heavy load.
Amazon Kinesis
New Features & Updates
2026/08/28 - Amazon Kinesis Data Streams announces streaming tables, delivering data to Apache Iceberg tables on Amazon S3 Tables
A fully serverless "streaming tables" feature has been added to Amazon Kinesis Data Streams that continuously delivers stream data to Amazon S3 Tables in Apache Iceberg format. This eliminates the need to build and operate your own Iceberg delivery pipeline, with data delivery costs reduced by up to 50%.
Inline intelligent compaction resolves the small file problem and stabilizes query performance, with downstream query costs reportedly reduced by up to 30%. It makes it easier to supply fresh data to near-real-time analytics and AI/ML feature pipelines.
2026/08/29 - Amazon Kinesis Data Streams announces data delivery to general purpose Amazon S3 buckets
Streaming data can now be delivered directly from Amazon Kinesis Data Streams to general-purpose Amazon S3 buckets. Delivery can be configured in just a few clicks from the console or API without chaining multiple serverless services or operating your own consumer applications on your own compute.
Batch analytics, log delivery, compliance storage, and replay can be achieved at data delivery costs up to 60% lower compared to custom implementations.
API Changes
2026/08/31 - Amazon Kinesis - 5 new 2 updated methods
Support has been added for data delivery from Amazon Kinesis Data Streams to Amazon S3 Tables (Apache Iceberg) and general-purpose Amazon S3 buckets, with the APIs CreateChannel, UpdateChannel, DeleteChannel, DescribeChannel, and ListChannels added.
Amazon MSK
New Features & Updates
2026/08/06 - Amazon MSK now delivers Kafka Authorizer Logs to customers
Kafka Authorizer logs can now be delivered at no additional cost for Amazon MSK Provisioned clusters (both Standard and Express brokers). The client IP address and the attempted API are recorded for each denied authorization request, making it easier to identify the cause of authorization errors on the client side.
Delivery destinations can be chosen from Amazon CloudWatch Logs, Amazon S3, and Amazon Data Firehose. It can be enabled from the MSK console or AWS CLI for both new and existing clusters, and is available in all regions where MSK Provisioned clusters are available, excluding the AWS European Sovereign Cloud.
2026/08/10 - Amazon MSK now supports in-place migration from Apache ZooKeeper to KRaft
Amazon MSK now supports in-place migration from Apache ZooKeeper to KRaft for existing clusters. Migration can be performed without provisioning new infrastructure, migrating data, or reconfiguring client applications, and the cluster remains available during migration.
Kafka 3.9.x is the last version to support ZooKeeper, and Kafka 4.x will support KRaft only, making this a practical migration path for upgrading to Kafka 4.x. For older versions, first upgrade to Kafka 3.9.x before starting the migration.
2026/08/17 - Amazon MSK now supports configuring custom domain names for MSK Provisioned clusters
Custom domain names can now be configured for Amazon MSK Provisioned clusters with either ZooKeeper or KRaft metadata management. Once a custom domain is defined per cluster, MSK automatically applies it to all brokers.
Previously, manual configuration was required for each broker, and it was not possible to configure at all for KRaft-based clusters. Since the configuration is maintained across scaling operations, client-side connection endpoints do not need to be changed during cluster migrations, DR failovers, or scaling.
2026/08/24 - MSK Replicator now supports OAuth 2.0 (SASL/OAUTHBEARER) authentication for replication from external Apache Kafka clusters to Amazon MSK
MSK Replicator now supports OAuth 2.0 (SASL/OAUTHBEARER) authentication when replicating from external Apache Kafka clusters to Amazon MSK Provisioned clusters. Target external clusters include on-premises, self-managed on AWS, and other clouds.
With OAuth/OIDC joining the previously supported SASL/SCRAM and mTLS, MSK Replicator can now be used for migrations from existing Kafka clusters that require OAuth authentication, DR with MSK as the failover target, and data distribution in hybrid and multi-cloud environments.
2026/08/31 - Amazon MSK Connect now supports restarting connectors
Newly created connectors and their tasks can now be restarted in Amazon MSK Connect. You can choose to restart the entire connector and all its tasks, or selectively restart only failed tasks.
Previously, recovering from failures required deleting and recreating the connector, causing downtime and operational overhead. Connectors without failed tasks can also be restarted, making it useful for recovery from transient issues and reflecting changes to external systems and dependencies. Available at no additional cost in all regions where MSK Connect is available.
API Changes
2026/08/06 - Managed Streaming for Kafka - 10 updated methods
MSK clusters can now output Authorizer logs alongside broker logs to a specified delivery destination.
2026/08/24 - Managed Streaming for Kafka - 2 updated methods
MSK Replicator now supports OAuth authentication when connecting to external Apache Kafka clusters, enabling replication from clusters that require OAuth for client authentication. Available in all regions where MSK Express brokers are available.
Amazon MWAA
New Features & Updates
2026/08/18 - Amazon MWAA Serverless now supports PythonOperator and BashOperator
Amazon MWAA Serverless now allows you to run custom Python functions and shell scripts directly on the serverless runtime using PythonOperator and BashOperator. Common code patterns such as data transformation, format conversion, and data quality checks can be executed without provisioning additional infrastructure.
You can bundle Python modules and shell scripts as a code package, upload it to Amazon S3, and reference it when creating or updating a workflow. The code at the time of workflow creation is fixed as a snapshot, and all subsequent executions use that snapshot, ensuring consistency across runs.
Amazon OpenSearch Serverless
New Features & Updates
2026/08/10 - Amazon OpenSearch Serverless now supports up to 10,000 collections per collection group
In the next-generation Amazon OpenSearch Serverless, the maximum number of collections per collection group has been raised from the previous 1,500 to 10,000. Collection groups bundle multiple collections together and allow OCUs (OpenSearch Compute Units) to be shared even among collections encrypted with different AWS KMS keys.
This reduces costs by enabling shared compute without provisioning individual OCUs per KMS key, and makes it easier to scale multi-tenant applications that provision collections per tenant. Per-collection security and access control are maintained.
Amazon OpenSearch Service
New Features & Updates
2026/08/06 - Amazon OpenSearch UI now supports Network Access Control
Network access control is now supported for OpenSearch UI applications. Using familiar IAM condition keys from other AWS services (aws:SourceVpce, aws:SourceVpc, aws:SourceIp), you can restrict access to only approved networks and configure consistent data boundaries across your entire environment.
Restrictions can be enforced at three layers: identity-based policies, VPC endpoint policies, and resource control policies (RCPs). Using RCPs, you can apply restrictions uniformly across all accounts in your organization, blocking users outside the network before they even reach the authentication screen.
2026/08/07 - Amazon OpenSearch Service announces additional upgrade runway for existing domains and support dates for additional versions
For the Extended Support for older versions announced in November 2024, security patches and OS patches have been extended by 12 months for customers who need more time to migrate, now through November 7, 2027. This applies to Elasticsearch 1.5–7.8 (excluding 5.6, which was already extended to November 7, 2028), OpenSearch 1.0–1.2, and OpenSearch 2.3–2.9.
Domains running affected versions will not be isolated. However, from November 7, 2026, the Extended Support surcharge for these versions will change to be equal to the instance price.
2026/08/17 - Amazon OpenSearch Service now supports automatic semantic enrichment for VPC domains
Automatic semantic enrichment has been extended to VPC-enabled domains. You can now use AI-powered semantic search without exposing your domain to the internet. Previously, this feature was only available for non-VPC domains.
Automatic semantic enrichment transforms keyword-only search into search that understands the meaning of a query. No self-management of machine learning models or custom integrations are required, and you can use it without modifying your existing VPC configuration.
2026/08/31 - Amazon OpenSearch Service adds new Cluster Insights for faster diagnosis of cluster status
17 new insights have been added to Cluster Insights that identify the root cause of Red/Yellow cluster statuses and provide specific recommendations for resolution. These cover resource exhaustion such as JVM OOM and CPU saturation, as well as configuration issues such as zone imbalance and misconfigured replica counts.
Previously, diagnosing the cause of Red/Yellow status due to unassigned shards required manually correlating multiple metrics across nodes and Availability Zones. The new insights, including 6 Critical ones that detect situations where primary shards become unassigned, suggest remediation steps such as scaling up instance types, adding disk capacity, and fixing shard allocation settings.
AWS Clean Rooms
New Features & Updates
2026/08/11 - AWS Clean Rooms supports exporting privacy-enhanced analysis logs for SQL
Privacy-enhanced analysis logs for SQL queries executed in AWS Clean Rooms collaborations can now be exported to an Amazon S3 bucket. You can view Spark execution details and use them to optimize queries and troubleshoot issues.
Export permissions are granted to members by the collaboration owner at the time of collaboration creation or via a change request to an existing collaboration. After a query is executed, logs can be output to any S3 path. For example, a third-party measurement provider can use this in a collaboration with a publisher to identify data skew that is causing slow queries.
2026/08/13 - AWS Clean Rooms supports minimum aggregation thresholds in custom analysis rules
Minimum aggregation thresholds can now be set in custom analysis rule types. This ensures that each row output by a query represents at least a specified number of unique values (such as user IDs), preventing results about individuals or small groups from being returned.
Data providers can specify their identifier column and the minimum number of identifiers to enforce on query output, and can set higher thresholds for specific columns. Previously, applying minimum aggregation thresholds to custom SQL required pre-approved analysis templates and manual code review.
API Changes
2026/08/11 - AWS Clean Rooms Service - 3 new 12 updated methods
AWS Clean Rooms now supports exporting obfuscated query execution logs.
2026/08/13 - AWS Clean Rooms Service - 8 updated methods
Support for minimum aggregation thresholds and comparison controls has been added to the custom analysis rule type.
In Closing
August 2026 was a month where updates centered around Apache Iceberg v3 and Amazon S3 Tables rolled out across various services. AWS Glue 6.0 and Amazon Redshift both added support for Iceberg v3, and managed integrations using S3 Tables as an output destination expanded significantly, including streaming tables in Kinesis Data Streams, long-term retention of system tables in Redshift, and metadata export from the Glue Data Catalog. Parts that previously required writing your own ETL pipelines can now be handled through configuration alone.
There were also developments on the governance front. Features including Amazon Quick's approval policies, default deny, and Microsoft Purview integration, trusted identity propagation for SageMaker notebooks, and Redshift's IAM Identity Center plus enhanced VPC routing support all point to a direction where per-user access control and auditability are assumed from the outset. If any of these updates catch your eye, please give them a try. We hope this article proves useful to someone.
