GitHub availability report: April 2026
In April, we experienced 10 incidents that resulted in degraded performance across GitHub services.
From latency to instant: Modernizing GitHub Issues navigation performance
How the GitHub Issues team used client-side caching, smart prefetching, and service workers to make navigation feel instant.
When "idle" isn't idle: how a Linux kernel optimization became a QUIC bug
We investigated a bug where CUBIC's congestion window became pinned at its minimum floor, causing a performance to plummet. The fix involved correctly measuring idle periods to distinguish RTT wait times from actual application idleness.
FigJam is now your coding agent’s whiteboard too
Agents are changing your code faster than your team can follow. Now you can close that gap with new MCP skills, architecture layouts, and more in FigJam.
Flow generation through natural language: An agentic modeling approach (2026) - Shopify
We fine-tuned Qwen3-32B into a tool-calling agent that generates Flow automations: 2.2x faster, 68% cheaper, with a weekly retraining flywheel.
Introducing Flagship: feature flags built for the age of AI
We are launching Flagship, a native feature flag service built on Cloudflare’s global network to eliminate the latency of third-party providers. By using KV and Durable Objects, Flagship allows for sub-millisecond flag evaluation.
Unweight: how we compressed an LLM 22% without sacrificing quality
Running LLMs across Cloudflare’s network requires us to be smarter and more efficient about GPU memory bandwidth. That’s why we developed Unweight, a lossless inference-time compression system that achieves up to a 22% model footprint reduction, so that we can deliver faster and cheaper inference than ever before.
Agents Week: network performance update
By migrating our request handling layer to a Rust-based architecture called FL2, Cloudflare has increased its performance lead to 60% of the world’s top networks. We use real-user measurements and TCP connection trimeans to ensure our data reflects the actual experience of people on the Internet.
Building the foundation for running extra-large language models
We built a custom technology stack to run fast large language models on Cloudflare’s infrastructure. This post explores the engineering trade-offs and technical optimizations required to make high-performance AI inference accessible.
GitHub availability report: March 2026
In March, we experienced four incidents that resulted in degraded performance across GitHub services.
How we optimized Dash's relevance judge with DSPy
We used DSPy to turn prompt engineering for our relevance judge into a measurable, automated optimization loop, improving task performance, cost, and how reliably it works in production.
Principles of Mechanical Sympathy
Modern hardware is remarkably fast, but software often fails to leverage it. Mechanical sympathy - a concept borrowed from racing and popularized in software by Martin Thompson - is the practice of creating software that is sympathetic to its underlying hardware. This practice can be distilled into a set of everyday principles: Predictable memory access, awareness of cache lines, the single-writer
The uphill climb of making diff lines performant
The path to better performance is often found in simplicity.
Our ongoing commitment to privacy for the 1.1.1.1 public DNS resolver
Eight years ago, we launched 1.1.1.1 to build a faster, more private Internet. Today, we’re sharing the results of our latest independent examination. The result: our privacy protections are working exactly as promised.
Making Turborepo 96% faster with agents, sandboxes, and humans
Turborepo 2.9 is up to 96% faster than Turborepo 2.8. Here's how we did it, using coding agents, a human in the loop, and Vercel Sandboxes.
Sandboxing AI agents, 100x faster
We’re introducing Dynamic Workers, which allow you to execute AI-generated code in secure, lightweight isolates. This approach is 100 times faster than traditional containers, enabling millisecond startup times for AI agent sandboxing.
Launching Cloudflare’s Gen 13 servers- trading cache for cores for 2x edge compute performance
Cloudflare’s Gen 13 servers double our compute throughput by rethinking the balance between cache and cores. Moving to high-core-count AMD EPYC ™ Turin CPUs, we traded large L3 cache for raw compute density. By running our new Rust-based FL2 stack, we completely mitigated the latency penalty to unlock twice the performance.
Introducing Custom Regions for precision data control
We are expanding Regional Services with new pre-defined regions and the launch of Custom Regions. Customers can now define precise geographical boundaries for data processing, tailored to meet their compliance and performance needs.
How we rebuilt the foundations of component instances
By replacing a decade-old architecture with a reactive foundation, we made common operations in large design systems up to 50% faster and unlocked a new way to build dynamic features.
Design’s influence is expanding, and here’s why that feels hard
As AI accelerates creation and raises expectations, designers are navigating a field that’s bigger, faster, and more demanding than ever.
Shopify’s journey to faster breadth-first GraphQL execution (2026) - Shopify
We questioned why conventional GraphQL execution incurs hidden costs, and rewrote it in a faster breadth-first manner to avoid them.
GitHub availability report: February 2026
In February, we experienced six incidents that resulted in degraded performance across GitHub services.
A QUICker SASE client: re-building Proxy Mode
By transitioning the Cloudflare One Client to use QUIC streams for Proxy Mode, we eliminated the overhead of user-space TCP stacks, resulting in a 2x increase in throughput and significant latency reduction for end users.
3 ways product teams are building conviction faster with Figma Make
Product managers at ServiceNow, Ticketmaster, and Affirm are using Figma Make to prototype their way forward.
How we rebuilt the search architecture for high availability in GitHub Enterprise Server
Here’s how we made the search experience better, faster, and more resilient for GHES customers.
Scaling redirects to infinity on Vercel
Redirects are trivial at a small scale, but at millions, latency and cost become real systems problems. This is the story of how Vercel implemented bulk redirects.
Can AI agents build real Stripe integrations? We built a benchmark to find out
State-of-the-art LLMs can solve many scoped coding tasks, but can they execute end-to-end software projects? To find out, we built the Stripe integration benchmark: an agentic test of real API integration work in a production-realistic environment.
Workflow lab: AI image tooling and interactive prototyping in Figma
A workflow to try—bringing together precise image editing, Vectorize, and interactive prototypes to move from concept to crit to completion faster.
How we rebuilt Next.js with AI in one week
One engineer used AI to rebuild Next.js on Vite in a week. vinext builds up to 4x faster, produces 57% smaller bundles, and deploys to Cloudflare Workers with a single command.
A Designer’s Guide To Eco-Friendly Interfaces — Smashing Magazine
Every high-resolution hero image, autoplay video, and complex JavaScript animation carries a cost. Sustainable UX challenges the era of “unlimited pixels” and reframes performance as responsibility. In 2026, truly sophisticated design is defined not by how much it adds, but by how thoughtfully it reduces its footprint.
GitHub availability report: January 2026
In January, we experienced two incidents that resulted in degraded performance across GitHub services.
Improve global upload performance with R2 Local Uploads
Local Uploads on R2 reduces request duration for uploads by up to 75%. It writes object data to a nearby location and asynchronously copies it to your bucket, all while data is available immediately.
7 learnings from Anders Hejlsberg: The architect behind C# and TypeScript
Anders Hejlsberg shares lessons from C# and TypeScript on fast feedback loops, scaling software, open source visibility, and building tools that last.
The three biggest agentic commerce trends from NRF 2026
The majority of retailers are actively implementing, or plan to implement, agentic commerce. They are refining their product catalog strategy to launch faster and investing in their own agentic shopping experiences.
GitHub Availability Report: December 2025
In December, we experienced five incidents that resulted in degraded performance across GitHub services.
Relative velocity and closing speed - Eli Bendersky's website
How To Measure The Impact Of Features — Smashing Magazine
Meet TARS — a simple, repeatable, and meaningful UX metric designed specifically to track the performance of product features. Upcoming part of the Measure UX & Design Impact (use the code 🎟 IMPACT to save 20% off today).
Comparing stats! PostgreSQL 18 against 17
This blog compares the stats of PostgreSQL 18 and PostgreSQL17. It is a great read with some common questions answered. Read to know more.
Stripe Atlas startups in 2025: Year in review
2025 was a breakout year for early-stage startups, as founders generated revenue faster than ever. Three shifts stand out: customer bases are becoming more global, time-to-revenue has compressed, and founders are turning their focus to AI agents.
Businesses grow revenue on Stripe 27 percentage points faster after accepting financing through Stripe Capital
In a new study, we found a strong causal relationship between accepting financing and growing revenue on Stripe. Learn which businesses are most likely to benefit, and how greater access to financing could drive significant GDP growth.
Figma Make is now available in Figma for Government
With Figma Make now available in Figma for Government, teams can prototype faster, collaborate securely, and accelerate progress on modernizing public services.
Let’s talk about GitHub Actions
A look at how we rebuilt GitHub Actions’ core architecture and shipped long-requested upgrades to improve performance, workflow flexibility, reliability, and everyday developer experience.
GitHub Availability Report: November 2025
In November, we experienced three incidents that resulted in degraded performance across GitHub services.
Python Workers redux: fast cold starts, packages, and a uv-first workflow
Recent advancements in Cloudflare Python Workers means fast cold starts, comprehensive package support, and a great developer experience. We explain how they were achieved and show how Python can be used to build serverless applications on Cloudflare.
From Chaos To Clarity: Simplifying Server Management With AI And Automation — Smashing Magazine
Server chaos doesn’t have to be the norm. AI-ready infrastructure and automation can bring clarity, performance, and focus back to your web work.
GitHub Availability Report: October 2025
In October, we experienced four incidents that resulted in degraded performance across GitHub services.
Figma opens a new hub in India to support our growing community of builders
With our new hub in Bengaluru, we’re investing in India’s fast-growing product development community.
Building world-class product search at Shopify: Where C++ excellence meets ML innovation (2025) - Shopify
Learn how we solved a major search engineering dilemma—running machine learning models at native C++ speed.
PostgreSQL 18: More performance with index skip scans
This blog describes the performance of Indexes in PostgreSQL version 18. You can read more to know how the performance matrix works.
StyleX: A Styling Library for CSS at Scale
StyleX is Meta’s styling system for large-scale applications. It combines the ergonomics of CSS-in-JS with the performance of static CSS, generating collision-free atomic CSS while allowing for expressive, type-safe style authoring. StyleX was open sourced at the end of 2023 and has since become the standard styling system across Meta products like Facebook, Instagram, WhatsApp, […]
Effectively Monitoring Web Performance — Smashing Magazine
There are lots of tips for [improving your website performance](https://www.debugbear.com/blog/improve-website-performance?utm_campaign=sm-10). But even if you follow all of the advice, are you able to maintain an optimized site? And are you targeting the right pages? Matt Zeunert outlines an effective strategy for web performance optimization and explains the roles that different types of data pl
Async QUIC and HTTP:3 made easy- tokio-quiche is now open-source
We’re excited to announce the open sourcing of tokio-quiche, our async QUIC library built on quiche and tokio. Relied upon in our services such as iCloud Private Relay and our next-generation Oxy-based proxies, tokio-quiche handles millions of HTTP/3 requests per second with low latency and high throughput.
Build better software to build software better
We manage the build pipeline that delivers Quip and Slack Canvas’s backend. A year ago, we were chasing exciting ideas to help engineers ship better code, faster. But we had one huge problem: builds took 60 minutes. With a build that slow, the whole pipeline gets less agile, and feedback doesn’t come to engineers until…
Counting Customers in PostgreSQL
This blog post focuses on realtime methods of counting in PostgreSQL. This includes a use case walkthrough. Read to know more.
So long, and thanks for all the fish- how to escape the Linux networking stack
Many products at Cloudflare aren’t possible without pushing the limits of network hardware and software to deliver improved performance, increased efficiency, or novel capabilities such as soft-unicast, our method for sharing IP subnets across data centers. Happily, most people do not need to know the intricacies of how your operating system handles network and Internet access in general. Yes, eve
Keeping the Internet fast and secure- introducing Merkle Tree Certificates
Cloudflare is launching an experiment with Chrome to evaluate fast, scalable, and quantum-ready Merkle Tree Certificates, all without degrading performance or changing WebPKI trust relationships.
Bun runtime on Vercel Functions
Vercel Functions now supports the Bun runtime, giving developers faster performance options and greater flexibility for optimizing JavaScript workloads.
A deep dive into BPF LPM trie performance and optimization
This post explores the performance of BPF LPM tries, a critical data structure used for IP matching. It delves into the trie algorithm, presents benchmark data, and explains how the current kernel implementation’s design choices lead to performance bottlenecks.
Benefits of a DESCending index
The DESC clause in CREATE INDEX is rarely needed. I'll show use cases for descending indexes, including storage efficiency and performance.
Unpacking Cloudflare Workers CPU Performance Benchmarks
Cloudflare investigated CPU performance benchmark results for Workers, uncovering and fixing issues in infrastructure, V8 garbage collection, and OpenNext optimizations. These improvements have made Cloudflare Workers faster for all customers.