All news in category “Vendor and Hyperscaler Watch”

🆕 AWS now supports Amazon EKS local clusters on first- and second-generation AWS Outposts racks that boot Amazon EC2 instances from EC2 instance store. This extends Outposts’ static stability benefits to EKS local clusters, keeping the entire Kubernetes control plane on the Outpost to meet data residency needs and reduce impact from temporary network disconnects. The updated architecture brings greater operational parity with cloud EKS, includes managed control plane responsibilities, and adds support for EKS add-ons and modern auth mechanisms.

🛠️ Amazon CloudWatch Application Signals now ranks service health on the application map and adds new infrastructure, logs, and traces tabs on the service overview page, enabling operators to triage unhealthy services and inspect compute, log snippets, and trace details in one place. These features surface runtime indicators for Amazon EKS, Amazon ECS, AWS Lambda, and Amazon EC2, and provide curated metrics with deep links to relevant monitoring tools to speed root-cause analysis. The capabilities are available in all Regions that support Application Signals.

🔒 Google Cloud announces collaboration with Apple to support Apple’s expanded Private Cloud Compute (PCC) systems on Google Cloud, built with Intel and NVIDIA. The effort leverages Google Cloud’s Titanium security architecture and Confidential Computing portfolio, including hardware Trusted Execution Environments, to protect data at rest, in transit, and in use. This layered approach aims to deliver verifiable integrity, no privileged runtime access, and enforceable privacy protections for sensitive AI workloads.

🔍 This post summarizes three business-focused takeaways from Microsoft Build 2026 for leaders evaluating AI adoption. It explains how Microsoft is shifting from standalone models to a shared enterprise intelligence layer—Microsoft IQ—that connects business data and processes across systems. The article highlights Azure’s agent platform and Foundry updates for production-grade deployment, governance, and performance, emphasizing that AI is expected to deliver measurable outcomes now.

🎯 Amazon Aurora PostgreSQL-Compatible Edition now supports PostgreSQL major version 18, starting with 18.3. This release brings community improvements to query performance and database management and adds the pg_roaringbitmap extension for fast, memory-efficient set operations on large integer collections. PostgreSQL 18 also introduces B-tree skip scans, retained optimizer statistics during major upgrades, and parallel streaming of large transactions for logical replication. Aurora 18.3 is available in all commercial AWS Regions and AWS GovCloud (US); upgrade options include RDS Blue/Green, in-place upgrades, or snapshot restores.

📣 Amazon Managed Workflows for Apache Airflow (MWAA) Serverless now publishes workflow and task state change events to Amazon EventBridge, enabling event-driven automation for Airflow pipelines. Previously, teams relied on custom polling or manual monitoring; this capability lets MWAA Serverless emit events for workflow states (started, running, succeeded, failed) and task states (scheduled, succeeded, failed, up for retry). Use cases include alerting on failures, triggering downstream pipelines when upstream workflows succeed, and logging state transitions for compliance.

🗂️ Elastic Beanstalk now surfaces Amazon CloudWatch Logs directly within the environment Logs tab, letting users view and filter log groups and streams without leaving the Elastic Beanstalk console. The Logs tab lists environment-associated log groups (including those matching the aws/elasticbeanstalk//* prefix), auto-selects the most recently active stream, and offers a stream dropdown and filtering options. A "View in CloudWatch" dropdown provides direct links to the log group, log stream, or CloudWatch Logs Insights for deeper analysis. This capability is available across all Elastic Beanstalk platform branches in AWS Commercial and AWS GovCloud (US) Regions where the service is offered.

🔍 Looker introduces dashboard agents in preview to enable conversational, in-dashboard exploration of BI data. The agent leverages the dashboard’s filters, cross-filters, and curated tiles to provide context-aware answers and can access underlying Explores for additional detail. Analysts can configure the agent with natural-language instructions to align responses with business logic, and the system surfaces intermediate reasoning, cited tiles, and applied filters to build user trust. Admins enable the feature in Looker 26.08.11+ via the Gemini settings.

🔐 AWS Lake Formation now allows reading and writing the underlying Amazon S3 data files for tables registered in the AWS Glue Data Catalog, unifying permissions for SQL and direct file access. It issues temporary, scoped credentials tied to Lake Formation table grants—SELECT for read and SUPER for read/write—and is supported in Amazon EMR 7.13+. You can use Spark or Trino with provided APIs or an open source plugin, and all activity is logged in AWS CloudTrail. This capability is available at no extra charge in supported Regions.

🔒 Check Point announced it has joined OpenAI’s Trusted Access for Cyber (TAC) program and the Daybreak initiative to access advanced cyber-capable models. The company will use GPT-5.5, OpenAI’s Codex agentic framework, and direct support from OpenAI to enhance threat analysis, incident investigation, detection engineering, and secure code review. Check Point emphasizes disciplined, focused application of these models to strengthen prevention, speed delivery, and maintain product security for enterprise customers.

🔍 Amazon Managed Service for Prometheus now supports ingestion, storage, and querying of Prometheus native histograms, enabling higher-resolution metric distributions with lower cardinality compared to classic histograms. DevOps and SRE teams can obtain more accurate percentile calculations without predefining bucket boundaries. Native histograms use exponential bucketing and store full distributions in a single time series, reducing series count and lowering costs by metering only populated buckets.

🔧 Amazon Managed Service for Prometheus now supports out-of-order sample ingestion and a workspace-level rule query offset. By default, workspaces accept late-arriving samples within a 1-minute window, adjustable or disableable. A global rule query offset delays rule evaluations so late samples can be included, improving alerting accuracy. Two new CloudWatch vended metrics provide visibility to tune these settings.

🔒 AWS announced the AWS Workload Credentials Provider, a lightweight client-side tool that automates export and deployment of certificates from AWS Certificate Manager and local caching of secrets from AWS Secrets Manager. It removes the need for custom EventBridge-based automation for certificate renewals, supports Windows and Linux, and works with Apache and NGINX. The provider is open source and compatible with Secrets Manager Agent functionality.

🔒 GitHub announced breaking changes shipping in npm v12 that disable install-time lifecycle scripts by default to reduce software supply chain risk. The update stops preinstall, install, and postinstall scripts, blocks resolving Git and remote URL dependencies unless explicitly allowed, and changes default allow settings such as --allow-git to "none". Developers are advised to upgrade to npm 11.16.0+, review warnings, and use npm approve-scripts to opt in for trusted packages.

🔒 GitHub will change npm v12 defaults in July to block automatic execution of dependency install scripts unless a project explicitly allows them. The change prevents preinstall, install and postinstall scripts — including implicit node-gyp rebuilds — from running by default, narrowing a common supply chain attack vector. Experts generally praised the move as overdue but warned it only reduces, not eliminates, supply chain risk and will push attackers to other methods.

🚀 AWS has expanded availability of OpenAI's GPT-5.4 and GPT-5.5 models to the US East (N. Virginia) Region on Amazon Bedrock. These models support a 272K-token context window, accept text and image input, and are accessible via the Responses API with server- and client-side tool calling, projects, and response streaming. GPT-5.5 targets advanced coding, research, analysis, and long-running agentic tasks, while GPT-5.4 focuses on frontier reasoning, coding, tool use, and long-context workflows.

🛠️ Amazon OpenSearch Service now integrates MCP Apps to bring observability workflows into compatible agentic IDEs like Claude Desktop and VS Code. With this integration, local AI agents can investigate incidents using logs, traces, metrics, and alerts stored in OpenSearch domains, collections, and Amazon Managed Service for Prometheus. Each MCP App returns both a concise text summary for agent reasoning and an interactive visualization for human review within the same conversation thread.

🔔 Amazon EC2 Future-dated Capacity Reservations now support cancellation, allowing customers to secure capacity up to 120 days in advance and cancel if plans change. Depending on timing, a cancellation charge may apply; when it does, EC2 provides a quoted charge for review and acceptance before cancellation completes. Cancellation charges are reportable in the AWS Cost and Usage Report (CUR) 2.0 under a new Capacity Reservation Status value. The feature is available to all Future-dated Capacity Reservation customers, with regional availability listed on the AWS Capabilities by Region site and further details in the Capacity Reservation user guide.

🔧 Amazon ECS Managed Daemons now support inter-task visibility and communication by adding pidMode and ipcMode settings. These allow daemons to share process and IPC namespaces with application containers when set to "shared", while the default "none" keeps them isolated. Daemons run once per instance and start before applications, enabling independent deployment of agents like tracing, profiling, and security tools. The feature is available in all AWS Regions at no extra cost.

⚡ Google Cloud has announced general availability of Lightning Engine for Managed Service for Apache Spark, available in both serverless and managed cluster modes. Built on Gluten and Velox with Google-engineered enhancements, the engine compiles Spark physical plans into native C++ to accelerate vectorized operations, window functions, and sorting. It also optimizes connectors for Cloud Storage and BigQuery and adds cost-based optimizations like hash table caching, aggregation pushdown, and adaptive shuffle partitioning.