Cloud Infrastructure
18 articles RSS
Broadcom's vSphere Express Targets Up to 128 Cores at a Reported $9,600 a Year; Analysts Differ on the Value of the EEA vSphere 8 Extension
Reports on Broadcom's vSphere Express cite up to 128 cores and a $9,600 figure that none of the cited reports attributes to Broadcom; analysts split on the EEA vSphere 8 terms.
Cloud Spend Limits Compared: Google Caps One Listed Service per Project, AWS Pauses a Whole Project in a Gradual Rollout, Azure Ties Limit to Credit
Google's spend caps, now generally available, pause one listed service in a project; AWS's limits pause a whole project in a gradual rollout; Azure's limit equals subscription credit.
Google Cloud Spend Caps Reach General Availability for a Limited Set of Services; Release Notes Say October 13, Google's Blog Says October 31
Google Cloud spend caps on Cloud Billing Budgets are generally available for four services, with an API and monthly resets; Google's release notes date it October 13 and its blog October 31, and the cited pages differ on which costs trigger a cap.
Amazon ECS Gateway Gives Each Service a Generated Hostname on Shared Application Load Balancers, and InfoQ Notes No Custom-Domain Guidance
AWS announced Amazon ECS Gateway on October 20, 2026. InfoQ reports neither the announcement nor the documentation describes how to put a custom domain in front of the generated hostnames.
Microsoft Sets Azure Deployment Environments to Retire Feb. 22, 2027 and Dev Box Sept. 18, 2028; ADE Has No Single One-to-One Replacement
Microsoft's retirement guides give February 22, 2027 for Azure Deployment Environments and September 18, 2028 for Dev Box; InfoQ notes the guides do not explain the withdrawals.
Kubernetes 1.37 Sets Timelines to Retire IPVS Mode and kube-dns Packages as EKS and GKE Offer Rollback and EKS Prices Extended Support
Sources on Kubernetes 1.37, EKS and GKE show retirement timelines for kube-proxy's IPVS mode and kube-dns packages, control-plane rollback on EKS and GKE, and EKS extended support at $0.60 per cluster-hour versus $0.10.
AWS Bahrain Data Loss, a Google London Zone Outage and AWS's Sovereign Cloud Test: What Commentators Say About Who Owns Cloud Resilience
After AWS said Bahrain data was unrecoverable, Azure gateways failed and a Google London zone went down, commentators say resilience must be designed and tested by customers.
GKE Sandbox Makes Its microVM Type Generally Available on Clusters Running 1.37.0-gke.4713000 or Later, With No GPU or TPU Support
Google's GKE release notes say the microVM sandbox type, built on Kata Containers and Cloud Hypervisor, became generally available October 7; Google's docs list nested virtualization as a requirement.
Kubernetes Blog Reports Up to 3x Pod Density From Local SSD Node Swap on GKE, With Gains of 25% to 200% Across Sandbox Runtimes
A Kubernetes Blog post reports that Local SSD-backed node swap raised pod density by 25% to 200% in tests of sandboxed browsers and Python runtimes, while project docs warn of noisy-neighbor risks.
Kubernetes Project Reminds Operators That the Kubelet Refuses to Start on cgroup v1 Nodes by Default Since v1.35
A Kubernetes Blog post says failCgroupV1 defaults to true since v1.35, so the kubelet won't start on cgroup v1 nodes unless an admin sets a temporary override.
Azure Gateway Outage Hit 18 Regions as a Gateway Manager Change Collided With OS Servicing, Microsoft Says
Microsoft says a gateway management service change, combined with unrelated OS servicing, disrupted Azure ExpressRoute and VPN gateways from September 30 to October 1; a post incident review is pending.
Uber Details ServiceScale, a Second Kubernetes Scaling Controller Built to Let Failover Reuse Idle Capacity
Uber describes a new ServiceScale controller that lets multiple orchestrators set replica counts on the same Kubernetes workloads, and the stale-cache and multi-writer bugs its year-long rollout exposed.