AWS announces a new cohort of AWS Heroes ahead of re:Invent, profiling community members across serverless, DevTools, and database work. The post is a recognition and community announcement with no technical instruction, architecture, or engineering detail for the reader to apply.
Amazon Web Services (AWS) engineering blog
Loading
AWS's weekly roundup highlights the generally available AWS Emulator for testing AWS API calls locally, and Nitro Sandboxes for VM-level isolation of EKS pods and ECS tasks. It also covers Cognito OAuth 2.0 token exchange, OpenSearch cross-Region disaster recovery objectives, and the new default AgentCore runtime. Most items are announcements, with some concrete details on recovery objectives, token claims, and fidelity limits.
AWS Emulator is a free, open-source local emulator written in Rust that lets developers run AWS CLI, SDK, and coding-agent workloads against a local HTTPS endpoint. The post explains its SigV4 enforcement, embedded mode for Rust test suites, and a fidelity-testing method that checks responses against real AWS. It also notes limits such as the lack of IAM policy evaluation.
AWS's weekly roundup covers recent launches, including the AgentCore Runtime V2 snapshot-based cold start design, managed Hive-to-Iceberg migration in S3 Tables, and DynamoDB bulk loading from S3. It also summarizes a MySQL binlog optimization for large transactions and Lambda queue mode for self-managed Kafka. Most items are product announcements, with a few concrete engineering details.
Amazon S3 Tables now supports managed migration of AWS Glue-registered Hive Parquet tables to Apache Iceberg. The workflow runs through Start, Sync, and Complete API calls, with a compatibility report generated before any data is copied, the destination read-only during migration, and write downtime limited to a brief cutover. The post also covers billing, supported regions, and the current scope limited to same-account, same-region tables.
AWS's weekly roundup covers Step Functions workflows exposed as AI agent tools through AgentCore Gateway, OpenAI's GPT-6.1 Terra on Bedrock, and several infrastructure launches including post-quantum ML-KEM in KMS, SOCI lazy loading for Lambda, and EKS Auto Mode Gateway API support. It also highlights customer case studies on DR automation, API-to-agent tooling, and S3 lifecycle cost reduction.
AWS Weekly Roundup for October 26, 2026 covers several launches, including AWS Continuum for code vulnerabilities, which uses AWS honeypot telemetry to argue that patch windows are shrinking. It also summarizes Amazon ECS Gateway, TorchNeuron for Trainium, Amazon Polly's Generative engine upgrade, and a zero-downtime ElastiCache migration path. Most entries are announcements with limited depth, though the vulnerability timeline and ECS Gateway design add some practical detail.
AWS announces general availability of Continuum for code vulnerabilities, a service that builds a context graph from code repositories and AWS environment data to confirm which findings are exploitable and rank them by attack-path proximity to critical systems. It validates findings with sandboxed exploit reproduction where possible and opens human-approved fix pull requests.
Amazon ECS Gateway is a new account-level resource that lets platform teams define shared ingress standards once and have ECS provision per-service ACM certificates, DNS hostnames, listener rules, and security groups automatically. Services can share an ALB (up to 25 per load balancer), and the post covers the many-to-many cluster associations, deployment-time compatibility rules, and capacity limits such as maxAlbCount.
AWS's weekly roundup covers Lambda Web Functions, which run standard HTTP servers such as Express and Next.js on Lambda without adapters and bill for CPU time rather than wall-clock time. It also covers S3 provisioning on the Vercel Marketplace using short-lived OIDC credentials, and an AWS Security Agent that exploited 33 of 37 recent CVEs. The post argues that agent limits should be enforced outside the agent's own reasoning.
Amazon S3, including S3 Tables and S3 Vectors, can now be provisioned from the Vercel Marketplace and v0 by Vercel. The post walks through connecting buckets to Vercel functions using short-lived OIDC credentials exchanged with AWS STS, querying Iceberg S3 Tables with DuckDB, and running vector queries against S3 Vectors. It is mainly a getting-started guide for the new integration.
The AWS Weekly Roundup for October 13, 2026 recaps Prime Day 2026 infrastructure metrics, including DynamoDB, Lambda, and Fault Injection Service usage, and covers recent launches such as Claude Haiku 5.5 on Bedrock, Nova 2.5 Sonic, GuardDuty RDS data-activity protection, Lambda OAuth for self-managed Kafka, and Redshift Iceberg materialized views. It also introduces Strands Box, an open source agent sandbox with policy-based controls. Most entries are brief announcements with limited...
AWS introduces Lambda Web Functions, an HTTP-native serverless runtime for standard Node.js frameworks such as Express, Fastify, Hono, and Next.js. The post explains multi-concurrent execution, active CPU billing, automated multi-region routing, revision-based canary deployments, and CloudFront integration.
AWS reports the infrastructure metrics behind Amazon Prime Day 2026, covering compute, storage, databases, caching, messaging, streaming, observability, and security at extreme scale. The post highlights service throughput, year-over-year growth, and 44,000 fault-injection experiments used to support availability during peak demand.
AWS’s weekly roundup covers Amazon Bedrock Managed Agents powered by OpenAI, new frontier models, the AWS Well-Architected Agent, Aurora PostgreSQL queries over Iceberg and Parquet data, and S3 Tables’ Iceberg V3 support. It also lists service lifecycle changes, Kiro workflows, and other AWS announcements.
AWS introduces the public preview of Well-Architected Agent, which analyzes AWS resources, utilization, topology, and business goals to generate prioritized recommendations across cost, security, performance, and resilience. It provides console, CLI, and infrastructure-as-code remediation guidance, plus API access for operational workflows.
Amazon Aurora PostgreSQL can now query live operational data alongside Apache Iceberg and Parquet data in S3 without ETL pipelines. The post explains DuckDB integration, foreign-table setup, Glue catalog federation, query optimizations, caching, and options for materializing hot data for lower latency.
Amazon S3 Tables now support the full Apache Iceberg V3 type system, including variant, geospatial, and nanosecond timestamp types, plus deletion vectors and row lineage. The post explains how to create or upgrade tables, query semi-structured data, build incremental pipelines, and account for engine and format compatibility.
AWS adds metadata pre-filtering to Amazon S3 Vectors, resolving filters before similarity search to improve recall by up to 5x on selective queries. The update supports operators such as $startsWith, up to 2 KB of metadata per vector, and in-place migration to enhanced indexes without re-ingestion.
AWS introduces three new Heroes from India, the United States, and China, recognizing their contributions to serverless computing, cloud education, technical content, and developer communities. The post highlights their community leadership, mentoring, and efforts to broaden access to AWS knowledge.