AWS Launches Low-Cost Burstable EC2 T8i Instances
AWS made Amazon EC2 T8i instances generally available on September 17, 2026, in four sizes — nano, micro, small, and medium — each offering 2 vCPUs with memory ranging from 0.5 GB to 4 GB and baseline CPU performance between 5% and 20% per vCPU. AWS says T8i delivers up to 30% better price-performance and up to 70% higher compute performance than the previous-generation T3 family, alongside improved network bandwidth.
The instances are available On-Demand and as Spot capacity, with Savings Plans support coming later; the two smallest sizes qualify for AWS Free Tier. T8i launched across 13 regions spanning US East/West, several Asia Pacific locations (Hyderabad, Malaysia, Mumbai, Seoul, Singapore, Sydney, Tokyo), Canada Central, and three European regions (Frankfurt, Ireland, London, Paris), with further regional expansion planned. AWS is targeting microservices, dev/test environments, low-traffic web workloads, batch data processing, and lightweight AI inference — the kind of intermittent-load workloads burstable instances have always suited, now at a meaningfully lower price point.
Read more — AWS News Blog
AWS Elastic Beanstalk Adds a Fully-Managed Cluster Mode on EKS
Also on September 17, AWS introduced Elastic Beanstalk Cluster Mode, a new deployment option that runs multiple applications on a shared, fully-managed Amazon EKS cluster instead of provisioning isolated infrastructure per application. Developers submit source code, a Dockerfile, or a container image in one of several supported languages (Java, .NET, Python, Node.js, PHP, Ruby, Go), and Beanstalk handles containerization and deployment onto the shared cluster automatically.
AWS manages the full lifecycle — deployment, scaling, patching, monitoring, and upgrades — and Cluster Mode adds AI-powered troubleshooting, traffic-splitting deployments with automatic rollback, event-driven autoscaling, and built-in OpenTelemetry observability. The pitch for teams running a portfolio of smaller applications is straightforward cost amortization: per-application cost drops as the portfolio grows since infrastructure is shared rather than duplicated, without the operational overhead of running EKS directly. Cluster Mode itself carries no additional charge beyond the underlying EKS control plane, compute, ECR, and CloudWatch usage, and it's available in all regions where Elastic Beanstalk operates.
Read more — AWS News Blog
Google Cloud Storage Intelligence Advisor Reaches General Availability
Google Cloud made Storage Intelligence Advisor generally available this month, adding zero-configuration monitoring across Cloud Storage projects, folders, and organizations. The advisor automatically baselines normal activity and flags anomalies — including surges in storage operations, unexpected increases in cross-region egress, and spikes in errors — without requiring any setup, and pairs each detection with prescriptive, actionable remediation steps rather than just a raw alert.
The feature also provides drill-down visibility into which specific resources are driving a detected change, letting teams trace a cost or performance anomaly back to the bucket or workload responsible before it turns into a larger bill or outage. For organizations running Cloud Storage at scale across many projects, automated cross-region egress detection in particular addresses a common blind spot — egress cost spikes are often only noticed after the billing cycle closes, and Storage Intelligence Advisor's baseline-and-alert approach catches them while they're still active.
Read more — Google Cloud Blog
Links & Sources
You May Also Enjoy
