OffNet Newsroom

Daily topic roundup

AWS

Tuesday, September 08, 2026 · 8 stories, curated & summarized — click any story for the source.

Share this roundup on LinkedIn Full edition
Hacker News (100+ points) general

Broadcom removes VDDK downloads, complicating VMware exit strategies

Broadcom has pulled access to the VMware Virtual Disk Development Kit (VDDK) downloads, removing a key tool for VM migration. This move increases friction for organizations attempting to leave the VMware ecosystem, as VDDK is often used to backup and restore virtual disks across platforms. The removal directly impacts workflows that rely on this SDK for data portability during vendor transitions.

  • VDDK downloads are no longer available, blocking a common migration path.
  • Exiting VMware is now more difficult without this disk access tool.
  • Teams must find alternative methods for VM backup and restoration.
  • Vendor lock-in risks increase as official migration aids disappear.
  • Review current migration plans for dependencies on VDDK immediately.
CHECKLISTAction Plan for VMware ExitAudit current workflows for VDDK dependenciesIdentify alternative tools for VM backupPlan restoration strategies without VDDKUpdate migration timelines for new frictionAssess increased vendor lock-in risks
Read at source →
The Register general

Thailand halts all datacenter construction and approvals pending review

Thai authorities have suspended all datacenter build permits and new project approvals to reassess regulatory frameworks. The pause applies broadly to existing and planned facilities, signaling a shift in how the nation manages digital infrastructure growth. This moratorium creates immediate uncertainty for operators seeking to expand capacity in the region.

  • All new datacenter construction permits in Thailand are currently on hold.
  • Existing approvals may be subject to review or cancellation under the pause.
  • Operators should expect delays in expansion plans and capital expenditure timing.
  • Regulatory scrutiny is intensifying, likely leading to stricter compliance requirements.
Read at source →
InfoQ generaldevops

Netflix opensource Flink autoscaler for 30k jobs, cutting costs 58%

Netflix is transitioning over 30,000 streaming jobs to an open-source Apache Flink Autoscaler across multiple AWS regions. This operator-level strategy resolves scaling inefficiencies found in their previous cluster-level approach for complex stateful pipelines. Early results show a 58% reduction in annual compute costs for one team, equating to $1.1 million in savings.

  • Operator-level autoscaling outperforms cluster-level methods for stateful Flink pipelines.
  • Deployment spans 30,000+ streaming jobs across multiple AWS regions.
  • One team achieved 58% compute cost reduction, saving $1.1M annually.
  • Netflix is releasing the solution as open-source for broader community use.
BY THE NUMBERS58% Cost Reduction with Open Source Autoscaler58%Annual compute cost savingsOne team saved $1.1 million yearly
Read at source →
AWS What's New awsdatabase ↺ since 09-04

Aurora MySQL Adds Multi-Source and Delayed Replication

Amazon Aurora MySQL now supports multi-source replication, allowing a single cluster to ingest data from multiple independent source databases simultaneously. This capability simplifies consolidation tasks such as merging shards or aggregating regional data for centralized reporting. Additionally, delayed replication enables configuring binlog replicas to lag behind their source by a specific duration, providing a safeguard against accidental data loss.

  • Consolidate multiple MySQL shards or regional instances into one central Aurora cluster for unified operations.
  • Use delayed replication as a safety net to recover from human errors or bad logic deployments.
  • Simplify backup and reporting workflows by aggregating data from disparate sources into a single destination.
  • Check MySQL multi-source replication documentation for configuration details and limitations.

Fleet impact: For Aurora MySQL fleets, multi-source replication simplifies the architecture for merging shards or aggregating regional data into a central reporting cluster without complex ETL pipelines. Delayed replication offers a straightforward mechanism to protect against logical errors or accidental deletes by maintaining a lagged replica for point-in-time recovery.

CHECKLISTAurora Multi-Source SetupConsolidate MySQL shards into one central clusterUse delayed replication as a safety netAggregate disparate sources for unified reportingReview documentation for configuration limits
Read at source →
AWS Database Blog awsdatabase ↺ since 09-05

CloudWatch Database Insights now monitors self-managed PostgreSQL on EC2

Amazon CloudWatch Database Insights has expanded its scope to include self-managed databases, specifically allowing monitoring of PostgreSQL instances running on Amazon EC2. This update enables engineers to view DB Load, Top SQL, and wait event analysis for self-managed workloads within the same console used for Amazon Aurora and Amazon RDS. The feature aims to unify observability across hybrid database fleets without requiring separate tooling.

  • Unified console for monitoring both managed RDS/Aurora and self-managed EC2 PostgreSQL instances.
  • Access to DB Load, Top SQL, and wait event analysis for self-managed workloads.
  • Reduces context switching by consolidating database observability into a single interface.
  • Simplifies fleet-wide performance tuning with consistent metrics across database types.
Read at source →
AWS Database Blog awsdatabase ↺ since 09-04

AWS Advanced JDBC Wrapper Aurora Global Database Write Forwarding Troubleshooting

Configuring the AWS Advanced JDBC Wrapper for Aurora Global Database write forwarding demands precise region-specific settings to avoid latency spikes and connection failures. The guidance covers correct dialects, plugin selections, host patterns, and write forwarding configurations for both primary and secondary regions. Misconfiguration in these areas directly impacts application stability and data consistency across the global topology.

  • Use region-specific settings in the JDBC wrapper to prevent latency spikes during cross-region writes.
  • Verify dialect and plugin configurations match the specific Aurora Global Database topology.
  • Check host patterns and write forwarding settings for both primary and secondary regions.
  • Ensure correct configuration to avoid connection failures when forwarding writes to secondary regions.

Fleet impact: For fleets running Aurora Global Database, this update highlights critical configuration dependencies for the Advanced JDBC Wrapper. DBAs must audit their wrapper settings for region-specific dialects and host patterns to prevent write forwarding failures and latency issues in multi-region deployments.

Read at source →
AWS What's New awsdatabase ↺ since 09-05

Bedrock Managed Knowledge Base now supports user-managed auth for SharePoint, OneDrive, Confluence

Amazon Bedrock Managed Knowledge Base introduces a user-managed setup flow for SharePoint, OneDrive, and Confluence sources. This shift replaces the previous requirement for admin-generated 2LO credentials with a 3LO flow where users sign in with their existing accounts. The service handles authentication automatically, significantly reducing setup time and removing the need for privileged access to third-party systems.

  • Eliminates dependency on admin-level credentials for third-party data sources
  • Reduces setup time for SharePoint, OneDrive, and Confluence integration to minutes
  • Enables non-admin users to prototype AI assistants grounded in corporate wikis and docs
  • Simplifies governance by handling authentication via user-managed 3LO flow
TRADE-OFFAuth Method ShiftOld 2LO MethodRequires admin credentialsComplex setup processPrivileged access neededNew 3LO MethodUser-managed sign-inMinutes to configureNo admin accessvs
Read at source →
AWS What's New awsdatabase ↺ since 09-05

AWS Bedrock Managed Knowledge Base adds native ServiceNow connector

Amazon Bedrock Managed Knowledge Base now includes a native connector for ServiceNow, eliminating the need for custom ingestion pipelines. Users can link their ServiceNow instance directly to ingest knowledge articles, service catalog items, and file attachments. The connector automatically handles data crawling, metadata extraction, and incremental syncs to keep the knowledge base current.

  • Native ServiceNow integration removes the overhead of building custom RAG ingestion pipelines
  • Automatically crawls knowledge articles and service catalog items for AI agent access
  • Supports incremental syncs to keep the managed knowledge base up to date
  • Handles metadata extraction and file attachments without manual intervention
HOW IT WORKSServiceNow to Bedrock Sync1Link ServiceNow instance2Crawl articles and items3Extract metadata4Sync to knowledge base
Read at source →
↑