April 2025 Summaries

The Model Context Protocol (MCP) is a communication protocol for AI agents to interact with services. The MCP specification has been updated to introduce a new transport mechanism, called Streamable HTTP, which simplifies how AI agents can interact with services by using a single HTTP endpoint for sending and receiving responses between the client and the server. This new transport method eliminates the need to manage separate endpoints for requests and responses, reduces complexity, and enables bi-directional communication between the server and clients. The MCP specification is rapidly evolving, and developers can start testing it today using tools like mcp-remote. With the introduction of Python language support in Cloudflare Workers, developers can now build MCP servers entirely written in Python, making it easier to bring their services into the AI ecosystem.

The massive power outage that struck Portugal and Spain on April 28, caused significant disruptions to transportation, retail businesses, and everyday activities. The electrical grid operator in Portugal initially blamed a fault in the Spanish electricity grid for the outage, but later denied these claims. Cloudflare's vast network and customer base enabled it to observe the impact of the power outage at both local and national levels, as well as across traffic, network quality, and routing metrics. In Portugal, internet traffic dropped significantly, with a 50% decrease within five hours, while in Spain, internet traffic fell by around 60%. The power outage also impacted network quality, with download speeds dropping to as low as 15 Mbps in Portugal and 19 Mbps in Spain, and latency increasing to as high as 50 ms in Portugal and 40 ms in Spain. Network infrastructure was also affected, with a drop in announced IP address space in both countries. The impact of the power outage was felt across Europe, with parts of France, Andorra, and Morocco also reportedly experiencing disruptions. As power is restored, internet traffic and other metrics are expected to recover.

Cloudflare has released its 21st edition of the Cloudflare DDoS Threat Report, which provides a comprehensive analysis of the evolving threat landscape of Distributed Denial of Service (DDoS) attacks based on data from the Cloudflare network. The report focuses on the first quarter of 2025 and highlights a significant increase in DDoS attacks, with 20.5 million attacks blocked in Q1, representing a 358% year-over-year (YoY) increase and a 198% quarter-over-quarter (QoQ) increase. Hyper-volumetric DDoS attacks, which exceed 1-2 Tbps or 1 Bpps, were also notable, with over 700 such attacks blocked in Q1. The report highlights emerging threats such as CLDAP reflection/amplification and ESP reflection/amplification attacks, as well as the importance of automation and proactive security strategies to stay ahead of DDoS threats. Cloudflare's global network with 348 Tbps of capacity remains dedicated to delivering unmetered, unlimited DDoS protection.

Cloudflare's network spans over 330 cities in 125 countries, providing a broad range of services to millions of customers. The company observes the impact of Internet disruptions at both local and national levels, as well as network level. In Q1 2025, Cloudflare observed significant Internet disruptions due to cable damage and power outages, including severe storms causing outages in Ireland and Réunion, an earthquake causing ongoing connectivity issues in Myanmar, a reported cyberattack on Russian networks, and fires taking a telecom provider in Haiti offline briefly. The company also saw widespread power outages in Angola, Sri Lanka, Chile, Honduras, Cuba, and Panama due to various causes such as storms, earthquakes, and fires. In contrast, Q4 2024 saw only one government-directed Internet shutdown, which is an unusual occurrence.

Cloudflare is a mission-driven company committed to building a better Internet by providing a comprehensive connectivity cloud platform that protects, connects, and accelerates operations. With record revenue growth, increasing traction among large customers, and an expanding suite of security solutions, Cloudflare is emerging as the partner of choice for enterprises and service providers across EMEA. The company is investing in a partner-first strategy to drive growth, with a focus on accelerating partner-led revenue, expanding market penetration, supporting partners to build service practices, and partnering in the best way possible. Joining this effort is an opportunity to assemble and lead a world-class partnerships team, fostering a high-performance culture, hiring the best talent, empowering growth, creating a performance culture, and winning together with partners.

Cloudflare has announced several new features and updates during its Developer Week 2025 event, including the introduction of AutoRAG for fully-managed retrieval-augmented generation on Cloudflare's global network, Cloudflare Workflows which is now generally available with production-ready durable execution, a toolkit for AI agents that includes new Agents SDK support for MCP clients, authentication/authorization/hibernation for MCP servers, and Durable Objects free tier. Additionally, the company has introduced Build global MySQL apps using Cloudflare Workers and Hyperdrive, Cloudflare Realtime and RealtimeKit which enables developers to build real-time audio and video apps in days with SDKs for various programming languages, and a new streaming ingestion service called Pipelines. Other notable announcements include the acquisition of Outerbase to expand database and agent developer experience capabilities, Cloudflare Secrets Store (Beta) which securely stores and manages account-level secrets, Workers Observability which powers up with General Availability of Workers Logs and new Query Builder, and Workers VPC which allows applications deployed on Cloudflare Workers to connect to legacy cloud infrastructure. These updates demonstrate Cloudflare's commitment to supporting developers in building smarter and shipping faster.

Cloudflare, a technology company with a powerful mission, transparent culture, and significant impact, has attracted Plushcap as its new member. The unique combination of these factors resonates with Plushcap's own values and leadership philosophy, which emphasizes open communication, trust, and collaboration. Joining Cloudflare feels like a natural progression in Plushcap's career, offering an opportunity to contribute to a scaling company that is committed to making a positive impact on the world. The company's culture and values align perfectly with Plushcap's beliefs, making this a significant and exciting new chapter.

The Cloudflare team has been working on improving the quality and usability of their Workers AI platform, a distributed inference platform that makes inference accessible to everyone. They have introduced several new features, including speculative decoding, which speeds up inference times by 2-4x, an asynchronous batch API for large workloads, and expanded LoRA support for more customized responses. The team has also updated pricing and added a new dashboard to improve the usability of the platform. Additionally, they have released new models, including four new ones, with improved performance and capabilities compared to their existing models. These updates aim to make Workers AI faster, more reliable, and more customizable while reducing costs.

Workers VPC is a new feature from Cloudflare that makes it easier to build cross-cloud apps by providing isolated environments for resources and allowing them to access each other securely. It introduces two new capabilities: Workers VPC, which groups resources into isolated environments, and Workers VPC Private Links, which connects a Workers VPC to an external cloud using either standard tunnels over IPsec or Cloudflare Network Interconnect. This allows developers to build secure, global, cross-cloud apps on Workers without exposing their private resources to the public Internet. The feature is designed to unlock access to external cloud resources and provide a dead-simple solution for managing complex architectures. It enables bidirectional connectivity between Cloudflare and external clouds, reducing egress fees and allowing developers to place resources wherever needed for optimal performance. Workers VPC is expected to roll out later in 2025 and will be built on the foundation of Cloudflare's existing private networking products.

Cloudflare has revamped its Startup Program to provide more flexible credit tiers, better aligning with how startups and developers actually consume Cloudflare. The new program now offers four credit tiers that cater to different stages of a startup's journey, from bootstrapped stealth startups to scaling startups. These tiers are designed to offer simplicity and clarity by aligning with where the startup is in its growth journey. With this expansion, Cloudflare aims to provide startups of all sizes with access to the right level of support and resources, exactly when they need them. The new credit tiers can be used for various purposes such as building AI applications, creating immersive real-time experiences, building durable multi-step applications, and more. Startups have reported positive experiences with Cloudflare's Developer Platform, citing its benefits in terms of ease of use, scalability, and efficiency. By joining the Cloudflare for Startups Program, startups can unlock a wide range of possibilities and build innovative products on top of Cloudflare's global network.

Cloudflare plays a crucial role in supporting the Internet's infrastructure as a reverse proxy for approximately 20% of all websites, improving performance, security, and reliability. The company collaborates with organizations like the National Center for Missing and Exploited Children (NCMEC) to combat child sexual abuse material (CSAM). Cloudflare Workflows was introduced to streamline the process of generating and submitting reports to NCMEC, providing a structured, observable, and retryable system that simplifies code behind it. The new architecture replaces a complex queue-based system with a single workflow that orchestrates the entire process from start to finish, offering real-time visibility into each report processed and the status of each step in the workflow. This transformation has improved debugging, error handling, and monitoring for NCMEC incident reports, making it easier to troubleshoot issues and retry failed steps or terminate workflows. Cloudflare Workflows can be deployed with minimal changes to wrap existing code within a Workflow, and the company provides documentation and starter projects to help developers get started.

The Certificate Transparency (CT) ecosystem is a critical component of the Web Public Key Infrastructure (WebPKI), ensuring that all website certificates are publicly visible and auditable. CT was developed in response to the 2011 DigiNotar hacking incident, which highlighted the need for a more transparent and accountable certificate issuance process. The system involves public certification authorities (CAs) issuing certificates that are then logged by CT logs, which are maintained by organizations such as Cloudflare. These logs provide a publicly visible audit trail of all certificates issued, helping to protect users and website owners from mis-issued certificates. Despite its successes, the CT ecosystem faces challenges in maintaining integrity and availability, particularly with regards to scaling and performance. To address these issues, a new static CT API log design has been developed, which provides efficiency improvements and reduces the risk of integrity violations. This design is now being deployed by organizations such as Cloudflare and Let's Encrypt, and is expected to play a critical role in ensuring the security and trustworthiness of the web.

Cloudflare Containers are a new platform that allows developers to run containers across the globe with little overhead, making it simple and scalable. Workers is an integration that enables users to extend routing and scheduling behavior as they see fit, allowing for fine-grained control over container lifecycle. Durable Objects provide programmable sidecars that can proxy requests to containers and manage their lifecycle. Cloudflare Containers are billed per 10ms of active compute time, with memory and CPU costs starting at $0.0000025 and $0.000020 per GB-second and vCPU-second respectively. In comparison, Google Cloud Run has similar pricing for Compute but charges more for Function instances. The platform aims to make it easy for developers to pick the right tool for the job, seamlessly shifting traffic between compute sources. The open beta is expected to be released in late June 2025.

Cloudflare has been instrumental in supporting the growth of startups by providing flexible infrastructure and innovative AI tools. Two notable examples are Lamatic AI and Skyward AI, both of which have leveraged Cloudflare's serverless computing capabilities to build scalable AI agent platforms. Lamatic AI uses Workers for Platform to deploy its AI agents, ensuring high availability, scalability, and security. Skyward AI has built an AI agent platform using Durable Objects and JavaScript-native RPC for AI coordination, but recently transitioned to Cloudflare's new AI agents framework. Both startups have achieved impressive results, including scaling to over three million serverless requests per month and delivering sub-100 ms response times for AI agent queries. Cloudflare's infrastructure has enabled these companies to simplify their architecture while supporting global availability without added operational complexity. The company is proud to see so many startups trust Cloudflare for a scalable, reliable, and efficient foundation.

Super Slurper is a data migration tool developed by Cloudflare that enables large-scale data transfers between cloud object storage providers and Cloudflare R2. The original architecture, built on top of Kubernetes, faced performance bottlenecks due to shared resources with other services and limited scalability. To address these issues, Cloudflare rearchitected Super Slurper using their Developer Platform, incorporating features like Cloudflare Workers, Durable Objects, and Queues. This new architecture improved transfer speeds by up to 5x, allowing for faster migrations of petabytes of data. The updated service now handles duplicate messages more effectively, ensuring each object is transferred only once. With its improvements, Super Slurper has become a crucial tool for developers to migrate their data from various cloud storage providers to R2.

D1 read replication makes read-only copies of your database available in multiple regions across Cloudflare’s network, lowering average latency by routing user requests to read replicas in nearby regions and increasing overall throughput by offloading read queries to read replicas. The new D1 Sessions API ensures sequential consistency for all queries in the session, regardless of which replica is used. This feature allows applications to serve read queries faster while maintaining strong consistency across requests, with automatically provisioned replicas in every region. The public beta is available now, and the feature will be refined and expanded based on developer feedback.

Apache Iceberg is quickly becoming the standard table format for querying large analytic datasets in object storage. It brings database-like features such as ACID transactions, time travel, and schema evolution to files stored in formats like Parquet or ORC. Historically, data lakes were just collections of raw files in object storage, but Iceberg solves these problems by providing a unified metadata layer that enables reliable, concurrent reads and writes, optimized metadata management, and schema evolution. With the launch of R2 Data Catalog in open beta, developers can now manage Apache Iceberg catalogs built directly into their Cloudflare R2 buckets, exposing a standard Iceberg REST catalog interface for connecting to existing engines like PyIceberg, Snowflake, and Spark. The catalog also ensures consistent, coordinated access to tables without conflicts or data corruption, making it an essential tool for data teams adopting Iceberg on R2.

The Cloudflare team is launching the open beta of Pipelines, a streaming ingestion product that allows users to ingest high volumes of structured, real-time data and load it into object storage service R2. The product enables scalable, fault-tolerant, and efficient data processing, with features such as automatic scaling, durable objects, and read-write replication for reliability. Arroyo, a cloud-native distributed stream processing engine, is being integrated with Pipelines to enable real-time data transformations and SQL-like querying capabilities. The service offers free tier plans and introduces pricing based on the volume of data ingested and delivered into Pipelines, with costs starting at $0.02 per additional GB beyond the first 50GB per month.

Cloudflare has introduced a new feature called Snippets, which allows users to control HTTP traffic using JavaScript "code rules" at no extra cost. Snippets are designed for performance, security, and control, providing a fast and flexible way to modify headers, rewrite URLs, retry failed requests, and more. The features include ultra-fast execution, granular request matching, sequential execution, native Cloudflare Rules integration, JavaScript and Web APIs support, automated deployment and versioning via Terraform, and no usage-based fees for Pro, Business, and Enterprise plans. Snippets are now generally available (GA) for all paid plans, with a new code editor to streamline writing and testing logic. The feature is ideal when built-in rule actions aren't quite enough, and it extends the model of Cloudflare Rules with greater flexibility for advanced scenarios.

Cloudflare has announced a new section called Workers Observability in their Cloudflare Dashboard, which allows users to query detailed log events across all Workers in their account. This feature provides deeper insights into system performance and can be used to answer questions such as "why" certain issues are occurring. The Query Builder tool enables users to construct structured queries to explore logs, extract metrics from logs, create graphical and tabular visualizations, and save queries for faster future investigations. Workers Observability also includes features such as CPU time and wall time monitoring, revamped Workers metrics, an Invocations view, and a Workers Observability API that allows users to programmatically retrieve their telemetry data. To use the feature, users need to enable Workers Logs in their Workers' settings or add a configuration to their wrangler file.

Cloudflare has launched a new service called Secrets Store in beta, designed to help customers centralize management of API tokens, keys, and credentials on the Cloudflare platform. The service aims to improve security by restricting access to sensitive values and making it easier for developers to manage secrets across multiple Workers scripts. With Secrets Store, users can create account-level secrets that are centrally managed and protected, using role-based access control (RBAC) to ensure only authorized personnel can view, create, edit, or delete secrets. The service also integrates with Cloudflare's existing products, such as Cloudflare Workers, AI Gateway, and Firewall Rules, allowing for seamless propagation of secrets across the network. While still in beta, Secrets Store is available for all customers via UI and API today, with plans to support more products and features in the future.

Cloudflare is introducing RealtimeKit, a suite of products that helps developers build real-time voice and video experiences. RealtimeKit combines Cloudflare's SFU, STUN, and TURN services with RealtimeKit, a collection of mobile SDKs and server-side services that abstract away the complexities of WebRTC development. The goal is to make it easier for developers to create interactive applications without needing extensive knowledge of WebRTC protocols. RealtimeKit focuses on developer experience, providing APIs tailored to common implementation patterns and handling device enumeration, permission requests, and UI rendering across platforms. The platform also integrates with AI providers like ElevenLabs, enabling seamless real-time AI communication. With RealtimeKit, developers can now implement complex real-time experiences in hours, not weeks, and are encouraged to try out the closed beta version for early access.

Cloudflare has announced that its network performance has improved, ranking as the fastest in 48% of the top 1000 networks by 95th percentile TCP connection time between November 2024 and March 2025, up from 44% in September 2024. The company uses Real User Monitoring (RUM) to collect data on its network performance, which is compared against other networks. Cloudflare's improvements are attributed to the presence of Edge Partner deployments, collaborations with ISPs, and investments in peering and traffic engineering. The company tracks its network performance to improve user experience and continues to optimize its traffic flow with each top 1000 eyeball network. With a focus on becoming the fastest network globally, Cloudflare aims to identify areas where it's slower, fix them, and share its updates with customers.

The Cloudflare team has announced support for MySQL in Cloudflare Workers and Hyperdrive, enabling developers to connect to their existing MySQL databases directly from Workers, without region restrictions. This is made possible by the Hyperdrive solution, which pools connections to databases globally, eliminates unnecessary roundtrips, and provides integrated caching to offload popular queries from databases. The team has also improved Node.js compatibility for the mysql package, allowing it to work on the Workers runtime, and has released a new version of the mysql2 package that supports static parsing without using eval(), making it compatible with Cloudflare Workers. To get started, developers can create new Hyperdrive configurations for MySQL databases hosted anywhere, using the Wrangler CLI or the dashboard, and configure their Wrangler configuration file to set up Node.js compatibility flags and specify the Hyperdrive binding. From there, they can use existing MySQL drivers with custom connection credentials provided by Hyperdrive, leveraging Cloudflare's network to speed up routing. The team is committed to iterating on support for MySQL during the beta phase and expanding the feature set of Hyperdrive to make it more flexible for full-stack workloads and performant for building global apps on Workers.

Hyperdrive is a service that enables developers to build performant global apps on Cloudflare Workers with their existing SQL databases. It reduces latency by caching query results and minimizing network round trips, allowing for faster and more secure connections between the Worker and the origin database. Hyperdrive uses Cloudflare's cache and architecture to achieve this, with a single global Endpoint for each Hyperdrive located near its origin database. This approach allows for significant reductions in latency and bandwidth usage, making it an attractive solution for developers building on Workers. The service is now available on the free plan of Cloudflare Workers, allowing developers to get started with a single Wrangler command or using the dashboard.

Cloudflare Workers now supports hosting, storing, and serving static assets for free, allowing developers to build full-stack applications within a single Worker. This feature was previously only available on Cloudflare Pages. The platform has also introduced several improvements to building apps on Cloudflare, including support for popular frameworks such as React Router v7, Astro, and Vue.js, as well as the ability to build complete full-stack apps without a framework using Vite and React together. Additionally, Workers now provides production-ready support for Node.js APIs, allowing developers to use libraries that rely on these modules, such as MongoDB, with minimal modifications. The platform has also raised the maximum CPU time per Worker request from 30 seconds to 5 minutes, allowing for compute-intensive operations to run for longer without timing out. Furthermore, Workers Builds now allows users to connect a Git repository to their Worker and have automatic builds and deployments on every pushed change, reducing setup and button-clicking needed to bring a project over.

You can now add a Deploy to Cloudflare button to the README of your Git repository containing a Workers application, making it simple for other developers to quickly set up and deploy your project. This feature creates a new Git repository on your GitHub/GitLab account, provisions resources needed by the app, configures Workers Builds (CI/CD), adds preview URLs to each pull request, and allows self-hosting with minimal setup. By adding this button, you can make your open-source project accessible to others without requiring deep infrastructure expertise, enabling them to deploy their own instance of your app in one shot, which means more traction, feedback, and contributions. The feature is available for both starter projects and full-stack applications, including complex pieces of software, making self-hosting easy with Cloudflare.

The Cloudflare Vite plugin brings Workers runtime directly into the Vite development server, allowing for a seamless match between dev and production behavior. This enables developers to iterate on client and server parts of their applications together without losing UI state between edits. The plugin also provides first-class support for React Router v7, enabling users to create new React SPA projects or update existing ones with ease. Additionally, the Cloudflare Vite integration supports full-stack frameworks like Astro, Qwik, SvelteKit, and others, drawing on Vite's development server, build pipeline, and phenomenal developer experience. The plugin unlocks many new possibilities for the ecosystem, including running server code in alternative JavaScript runtimes, such as workerd. With this release, developers can now develop web applications that are also cloud-native, using the full Cloudflare Developer Platform, including KV, D1, Service Bindings, RPC, Durable Objects, Workflows, Workers AI, etc.

The Cloudflare adapter for OpenNext is now available in beta version 1.0.0-beta, allowing users to deploy Next.js applications on the Cloudflare platform with improved integration and support for more features than its original version. The adapter is tightly integrated with OpenNext and supports many Next.js features, including cache handler, PPR, middleware, App router, Pages routers, and Cloudflare Images. However, some limitations still exist, such as Microsoft Windows support and Edge runtime support, which are planned to be addressed in future releases. The ecosystem has also seen updates, including improved NodeJS compatibility for Workers and increased Worker size limits. With the release of 1.0-beta, most Next.js 14 and 15 applications are expected to run seamlessly on Cloudflare, and a v2 release is planned with a focus on reducing bundle size, improving application performance, and allowing users to deploy to multiple Workers.

The Outerbase team has been acquired by Cloudflare, a leading infrastructure platform for developers. The acquisition aims to integrate Outerbase's database management tools into Cloudflare's platform, making data more accessible and intuitive for developers. Outerbase was founded on the idea of simplifying databases for everyone, regardless of their skill level, and its technology is now being merged with Cloudflare's Durable Objects and D1 services. The integration will bring features like a unified data explorer, new REST APIs, query editor with type-ahead functionality, real-time data capture, and more to Cloudflare's platform. The acquisition also marks the end of Outerbase's hosted cloud service, but self-hosting options will still be available. The Outerbase team is excited to join forces with Cloudflare and contribute to building a better developer experience for data management.

The Cloudflare Workflows platform is now Generally Available (GA), offering a durable execution engine that allows developers to build long-running, multi-step applications. It provides automatic retries and state persistence, making it ideal for building reliable distributed applications that can handle real-world failures gracefully. The new `waitForEvent` API enables workflows to wait for events in the real world, such as human approval or incoming webhooks, allowing for more dynamic and interactive workflows. Pricing has been adjusted to include storage pricing, with a 1 GB included per month, and additional storage costs at $0.20/GB-month. Users can start building with Workflows by visiting the documentation, reviewing the starter project code, and deploying it to their own Cloudflare account.

AutoRAG is a fully managed Retrieval-Augmented Generation (RAG) pipeline powered by Cloudflare, designed to simplify how developers integrate context-aware AI into their applications. It removes the complexity of building and maintaining a RAG pipeline, allowing users to focus on building smarter and faster applications. AutoRAG uses Workers AI's vector database, embedding model, LLMs, and custom indexing, retrieval, and generation logic to deliver a fully-managed RAG pipeline end-to-end. The pipeline continuously monitors data sources and indexes in the background, ensuring that the AI stays fresh without manual effort. Users can create an AutoRAG instance and monitor its indexing process, then integrate it into their application using the AI binding. During the open beta, AutoRAG is free to enable, and users are limited to 10 instances with up to 100,000 files per instance. The platform plans to expand data source integrations, improve response quality, and introduce new features throughout 2025.

Cloudflare has announced several updates to its Agents SDK, a platform for building AI agents. The new features include the ability for AI agents to connect to external services through MCP (Model Context Protocol), which allows them to interact with tools and resources served by different service providers. This enables users to give their AI agent secure access to their data and applications without having to build authentication flow from scratch. Additionally, Cloudflare is integrating its Agents SDK with Stytch, Auth0, and WorkOS to make it easier for developers to configure authentication and authorization for their MCP servers. The new updates also include support for hibernation in McpAgent instances, which allows stateful MCP servers to sleep during inactive periods and wake up with their state preserved when needed. Furthermore, Durable Objects are now available on the free tier, allowing developers to build AI agents on Cloudflare without committing to a paid plan.

Cloudflare has partnered with Meta to make the latest and most powerful model, Llama 4, available on their Cloudflare Workers AI platform. Llama 4 is an industry-leading release that combines a Mixture of Experts architecture with an early-fusion backbone, making it natively multimodal. The model consists of two components: Llama 4 Scout and Llama 4 Maverick, with the former being available on Workers AI today. Llama 4 Scout has a context window of up to 10 million tokens, which is one of the largest available in an open-source model. Despite having a large number of parameters, the MoE architecture can intelligently use only a fraction of them during active inference, delivering faster responses. The Mixture of Experts model works by having individual specialized neural networks called "experts" that work together to provide deeper results and faster inference times. Cloudflare Workers AI provides an efficient way to host Llama 4 models without worrying about infrastructure or hardware requirements.

Cloudflare is kicking off its 2025 Developer Week with a focus on innovation and announcements for developers. The past two years have seen rapid technological advancements, including the release of state-of-the-art models like ChatGPT and o4 model Ghiblify the world. This has made it easier for developers to experiment and ship code, expanding the audience of who can write code. However, a platform is needed to make deployment and scaling as easy as writing code. Cloudflare aims to provide a comprehensive platform that just works, offering building blocks such as compute, storage, databases, and AI. The company has been chipping away at adding more primitives to its platform over the past 8 years, with tools for local development, CI/CD, and observability. This Developer Week will see the release of new tools and products to continue ensuring developers have everything they need when building applications. Cloudflare is also committed to investing in the developer experience, listening to feedback and improving the platform based on it. The company invites developers to share their thoughts and feedback throughout the week.

The United States Cybersecurity and Infrastructure Agency (CISA) has released best practices for technology companies, known as the Secure-by-Design pledge. Cloudflare has signed this pledge to demonstrate its commitment to creating resilient systems where security is a foundational principle. The company's Linux kernel updates pose significant challenges due to the need to balance security patching with user experience. To address this, Cloudflare employs a calculated approach, carefully removing traffic from servers before rebooting them. This process involves marking servers for maintenance, stopping traffic, and disabling internal traffic. Additionally, the company has developed Reboau, an internally-built tool to manage custom reboot logic in its Control Plane. By leveraging Reboau and following CISA's Secure by Design principles, Cloudflare aims to improve security and reliability for its customers.

The Images binding is a new feature introduced by Cloudflare to connect its Images API directly to Workers, enabling developers to transform, overlay, and encode images within the Cloudflare Developer Platform ecosystem. This binding removes unnecessary friction from application development by allowing developers to work more productively with images. It supports functions such as .transform(), .draw(), .output(), and .info() that can be used to manipulate images. The binding works by establishing a communication channel between a Worker and its backend services, making it easier for developers to build scalable and cost-effective image pipelines. Additionally, the binding is available in local development mode without incurring any usage charges, allowing developers to test and debug their applications more efficiently.

The author of the text was trying to run a test suite for their internal Cloudflare instance using `yarn test`, but it kept hanging and eventually crashing after approximately 27 seconds. The problem was caused by Jest, a JavaScript testing framework, running a command called `sl` (which is actually a fun little joke program from 1993) multiple times in parallel. The `sl` command is responsible for creating an animated steam locomotive that chugs through the terminal. The author noticed that the `sl` command was being run 4 times concurrently by Jest, which took around 6.7 seconds to complete each time. This meant that the total execution time of the 4 concurrent commands would be around 27 seconds, causing the test suite to hang and eventually crash. The issue was resolved when the author renamed the `sl` command to something else, effectively removing the naming clash between Jest's `sl` and the joke program.

Cloudflare has announced that they are making all purge methods available to all customers, regardless of plan type. This includes methods such as purge by URL, purge by hostname, purge by tag, purge by prefix, and purge everything. The company has spent two years rebuilding its purge pipeline with a focus on increasing throughput while maintaining near-instant invalidation across its global network. The new Instant Purge system allows for consistent, rapid propagation of purge requests across the vast network, reducing storage requirements by 10x. Cloudflare supports multiple targeted purge methods to help customers achieve a balance between precision and speed, and has raised the default purge rate limits for Enterprise customers.