Infrastructure News

Vercel now allows anyone to submit their own services to its Vercel Connect integration directory. This change lets developers connect more tools to their projects, speeding up workflows without waiting for official Vercel-built connectors.

Spotify experienced a major outage affecting thousands of users, primarily on its mobile app. The incident serves as a critical reminder of the complex dependencies and potential single points of failure in modern cloud-based services.

Cloudflare has added native support for the HTTP Vary header. This gives developers more control over how content is cached, helping to reduce server load and prevent inefficient cache usage known as "cache thrashing."

Adobe has open-sourced a new tool that lets teams safely view their own GPU performance metrics in shared Kubernetes clusters. This solves a major security challenge by preventing teams from seeing each other's sensitive operational data.

Atlassian replaced its core metrics system, handling data from 100,000 servers, with OpenTelemetry. The high-stakes migration was done live without disrupting critical production alerts, proving the standard's maturity.

Google has added an AI assistant called 'Ask' to its Earth Engine platform. Powered by Gemini, it helps developers write complex geospatial code using simple text prompts, accelerating projects in climate science, agriculture, and urban planning.

Google Cloud has launched its new Z4D machine series, designed to speed up applications that handle large amounts of data. These new virtual machines and bare-metal servers target databases, analytics, and other I/O-intensive business-critical workloads.

Meta built a new gateway that reduced persistent database connections by 19 times. Called ZGateway, the stateless proxy handles over a billion operations per second, centralizing traffic routing, caching, and load balancing for its ZippyDB system.

Uber developed a new Kubernetes tool that separates scaling decisions from the actual scaling actions. This allows multiple systems to manage capacity safely, enabling regional failover without paying for constantly running idle servers.

The Java ecosystem saw a wave of updates, including major releases for TornadoVM and Groovy, plus new versions of GraalVM, Hibernate, and Quarkus. These releases offer developers performance gains, new features, and critical stability improvements across the board.

Google's new GKE Pod Snapshots can load a 70-billion-parameter AI model in just 37 seconds. The feature cuts startup latency by up to 89% by saving a model's memory state to Cloud Storage for faster reloads.

Fujifilm has released new LTO-10 data tapes, offering a massive 40TB of native storage capacity. This technology provides a cost-effective solution for long-term archival and backup for large enterprises managing huge datasets.

Docker has launched Cloud Sandboxes, secure hosted environments for running code. The new platform aims to provide a consistent experience for developers moving workloads from their laptops to the cloud, with a special focus on AI agents.

Cloudflare has replaced WordPress for its main blog, migrating to a new open-source system it built in-house. The new platform, EmDash, is designed for extreme performance and can handle up to 7,000 requests per second.

AI agents for coding often fail because they only remember conversations, not their work environment. A new approach gives them a persistent "workspace" to manage files and dependencies, letting them work just like a human developer.

AWS has updated its Model Context Protocol to be stateless, eliminating the need for complex "sticky sessions." This change simplifies server architecture, making it easier to scale applications horizontally and improve overall system resilience.

AI search company Perplexity replaced Amazon's DynamoDB with its own custom database written in Rust. The move cut query latency by five times and significantly reduced cloud storage costs, improving efficiency for its high-volume search infrastructure.

AWS has enhanced Amazon EventBridge, allowing a central event bus to receive events from other accounts. This update simplifies complex architectures for large organizations, reducing operational overhead and costs by centralizing event management and guaranteeing message order.

A popular technique for improving AI chatbots, called query decomposition, can actually make them less accurate. By breaking down complex questions, the system can lose the original context, leading to worse answers for users.

Google Cloud's 'fluid compute' strategy helps teams avoid AI hardware shortages. It lets developers design workloads that can flexibly run on different available accelerators, like GPUs or TPUs, preventing costly project delays.

Confidential AI lets companies use powerful AI models on sensitive information like patient records or financial data without ever exposing it. This technology could unlock new AI applications by solving the core problem of data privacy and control.

Cloudflare rewrote its DNS cache in Rust, freeing up 100 terabytes of memory across its network. The change also boosted performance, cutting lookup times by 19% and allowing for a larger cache without new hardware.

Fitness tracker company WHOOP was drowning in security alerts. They built a system to automate filtering, letting engineers focus only on critical threats while still having the final say on fixes, tackling widespread alert fatigue.

Modal rebuilt its entire infrastructure to support millions of concurrent sandboxes, launching tens of thousands per second. Their custom solution offers a new blueprint for massive-scale computing beyond traditional tools like Kubernetes.

A new modular architecture for Cloudflare Workers helps developers build safer multi-tenant SaaS apps. By breaking up monolithic workers into smaller services, teams can reduce deployment risks and limit the impact of failures for their customers.

The Kubernetes team managing core application tools, SIG Apps, is shifting focus to handle more complex workloads like AI and databases. This signals future changes to fundamental tools like Deployments, aiming for better lifecycle management for all users.

AWS launched CloudWatch Omni, a new tool to help developers understand why their AI agents behave unpredictably. It unifies monitoring to explain agent actions, a task traditional tools like the original CloudWatch have struggled with.

A startup veteran shared a lean architectural pattern using GCP, Firebase, and Cloud Run. This stack helps small teams build scalable products quickly, manage state efficiently, and maintain lean DevOps practices to accelerate product-market fit.

Changesets, a popular tool for managing JavaScript projects, released version 3. It's now 88% smaller, uses modern ESM-only code, and fixes long-standing issues with how it handles related software packages, making it faster and more reliable.

Vercel now displays billable duration and CPU minutes for each deployment. This gives developers and companies direct insight into their build costs, helping them optimize usage and manage their budget more effectively.