News For Dev-ops

It has been an absolutely outstanding year of content for .NET from creators around the globe sharing their passion for .NET and the .NET team giving insight into the latest and greatest in the world of .NET. From events, live streams, and plenty of on-demand content dropping on the .NET YouTube nearly every single day, it is a great way to stay up to date, but also get involved and give feedback to the team in real-time. In 2025, the .NET YouTube channel welcomed 31.2K new subscribers (346K total), racked up 5.3M views, and reached 533.6K hours of watch time. Along the way, we posted 169 long-form videos (including 75 .NET Conf sessions), and hosted 149 live streams. That’s over 22,000 days of watch time - about 60 years! Thank you for watching, engaging, and subscribing ❤️ Top .NET videos of 2025 It was fun looking back at this year's top videos as it really was a wide range of content. Evergreen favorites (watched in 2025) Some of the channel’s biggest hits in 2025 were ev...

A customer was using the v2 task_sequencer class we developed some time ago. (Here's the v1 task sequencer .) They found that they occasionally suffered from stack overflow crashes. QueueTaskAsync::<lambda_2>::operator()+0x714 std::coroutine_handle<void>::resume+0x6c task_sequencer::chained_task::complete+0x88 task_sequencer::completer::~completer+0x58 QueueTaskAsync::<lambda_2>::operator()+0xaf8 std::coroutine_handle<void>::resume+0x6c task_sequencer::chained_task::complete+0x88 task_sequencer::completer::~completer+0x58 QueueTaskAsync::<lambda_2>::operator()+0xaf8 std::coroutine_handle<void>::resume+0x6c task_sequencer::chained_task::complete+0x88 task_sequencer::completer::~completer+0x58 QueueTaskAsync::<lambda_2>::operator()+0xaf8 std::coroutine_handle<void>::resume+0x6c task_sequencer::chained_task::complete+0x88 task_sequencer::completer::~completer+0x58 QueueTaskAsync::<lambda_2>::operator()+0xaf8 std::coroutine_handl...

2025 was a big year for .NET developers. We kicked things off with early .NET 10 previews, kept momentum with steady tooling and runtime improvements, and wrapped it all up with the launch of .NET 10 in November. 🚀 .NET 10 .NET 10 was the headline event of the year: the latest Long Term Support (LTS) release with improvements across the runtime, SDK, libraries, languages, workloads, and tooling. Start here: Announcing .NET 10 - The launch post and a map to what’s new across the stack. See how it began: .NET 10 Preview 1 is now available! - Early themes across runtime, libraries, SDK, and workloads. ⚡ Performance Performance posts are one of the best ways to learn what really changed, because they come with benchmarks and code you can try yourself. Performance Improvements in .NET 10 - A deep tour through hundreds of improvements. Preparing for the .NET 10 GC (DATAS) - A focused read for folks tracking GC behavior and changes. 🤖 AI AI reshaped day-to-day .NET w...

Today, AWS launches simplified import of CloudTrail Lake data in Amazon CloudWatch, a data management and analytics service that allows you to unify operational, security, and compliance data across your AWS environment and third-party sources. With this launch, you can now import your historical CloudTrail Lake data into CloudWatch with a few steps enabling you to easily consolidate operational, security, and compliance data in one place. In CloudWatch, you simply specify the CloudTrail Lake event data store (EDS), and the date range to initiate import of your CloudTrail data. Simplified import of CloudTrail Lake data is supported via the AWS console, CLI, and SDK. While simplified import of CloudTrail Lake data is available at no additional cost, you incur CloudWatch fees based on custom logs pricing. To learn more about simplified import of CloudTrail Lake data and supported AWS regions, visit the Amazon CloudWatch documentation. Post Updated on December 30, 2025 at 02:00PM

A little while ago, I wrote about the use of color-keyed overlays to render video smoothly . The idea is that the CPU itself produces an image with a block of solid color (green in my example), and the graphics card is instructed to replace those pixels with the pixels from an off-screen image that it generated separately. Rather than doing the image composition in the CPU, the composition happens in the video card as the image leaves the card and goes to the monitor. In the subsequent discussion, some people remembered this technique but noted that in their recollection, the color-key was not green but some other color. What's going on? When you set up the color-keying with the video card, you gave it a few instructions. You told it where to find the off-screen replacement image and how big it was. You told it the size and location of the on-screen rectangle where the replacement image should go. And you told it the magic color to look for inside that rectangle. So the CPU got ...

Amazon OpenSearch Service now supports AWS KMS customer managed keys (CMKs) and increased size for OpenSearch UI metadata. Amazon OpenSearch UI is a managed service for dashboarding and operational analytics that provides a unified view across multiple data sources, including OpenSearch domains and collections, Amazon S3, Amazon CloudWatch, and AWS Security Lake. You can now create new OpenSearch UI applications with metadata encrypted with your own CMKs, helping organizations meet regulatory and compliance requirements. This launch also increases the metadata size limit for saved objects in OpenSearch UI, enabling you to create and store complex queries, extensive visualizations, and large-scale dashboards. CMK support and increased metadata size are available in all regions that OpenSearch UI is available . Learn more at Amazon OpenSearch UI Developer Guide . Post Updated on December 29, 2025 at 08:00AM

Amazon Connect dashboards now support filtering metrics based on custom business dimensions such as business divisions, product lines, or customer segments. Using predefined attributes, you can create business dimensions to filter metrics helping you customize the dashboards based on your unique business need. For example, if your queue handles contacts across product lines, you can filter metrics by product line to compare handle times and determine where agents need product training. Amazon Connect dashboards are available in all AWS commercial and AWS GovCloud (US-West) Regions where Amazon Connect is offered. To learn more about dashboards, see the Amazon Connect Administrator Guide . To learn more about Amazon Connect, the AWS cloud-based contact center, please visit the Amazon Connect website . Post Updated on December 29, 2025 at 03:00PM

Amazon Connect now automates agent performance evaluations in Portuguese, French, Italian, German, and Spanish using generative AI. Managers define custom evaluation criteria in natural language and receive AI-generated evaluations with justifications in their preferred language. Performance evaluations also supports cross-language evaluation and can complete assessments in English, even when the conversation is in another language. This enables multilingual contact centers to use a standardized evaluation framework across languages. This feature is supported in 8 AWS regions including US East (N. Virginia), US West (Oregon), Europe (Frankfurt), Europe (London), Canada (Central), Asia Pacific (Sydney), Asia Pacific (Tokyo), and Asia Pacific (Singapore). For information about Amazon Connect pricing, please visit our pricing page . To learn more, please visit our documentation and our webpage . Post Updated on December 26, 2025 at 06:00PM

Commenter Shawn wondered why we are so worried about memory access semantics. Back in my day, we just used full barriers everywhere, and I didn't hear nobody complainin' . Moore's law says that the number of transistors in an integrated circuit doubles about every two years. For a long time, these resources were employed to make CPUs faster: Accelerating clock speeds, expanding caches, deepening pipelines, performing more prediction and speculation, increasing the number of registers, that sort of thing. All of these improvements sped up single-threaded operations. The problem is that you can make things only so small before you start running into problems. It's hard to build things that are smaller than atoms. The speed of light constrains how quickly two components can communicate with each other. You have to look for other ways to get your performance gains. CPU speeds have largely leveled off at the 4 GHz range. You can't go faster that way any more. Instea...

Just to see what happens, a customer called Set­ClipboardData( CF_TEXT) and passed a memory block that was intentionally not null-terminated. They found that when they subsequently pasted the string (say, into Notepad), the last character didn't get pasted. As a second experiment, they over-allocated the buffer by one byte, still refused to null-terminate the string, and they found that all of the characters made it. There's nothing in the documentation for Set­Clipboard­Data that contains any warning about trying to put non-null-terminated strings on the clipboard. What's going on? While it's true that there nothing in the Set­Clipboard­Data documentation, there is text in the Standard Clipboard Formats documentation. CF_TEXT 1 Text format. Each line ends with a carriage return/linefeed (CR-LF) combination. A null character signals the end of the data. Use this format for ANSI text. The documentation says that you need to pasa a null-terminated strin...

A customer didn't understand why their Win32 accelerator key for Ctrl + M was triggering spuriously. Specifically, it was triggering when the user hit the Enter key, which is nothing like the Ctrl + M two-key combination. They defined their accelerator table like this: IDA_MAIN ACCELERATORS BEGIN "^M", IDM_MUMBLE END Accelerator key definitions can be done by character or by virtual key code. If you use a quoted string, then you are defining a character accelerator which triggers when that character is entered by whatever means. For example, if you define a character accelerator for "0" , it will trigger if the user presses the 0 key on the top row of the keyboard, or if they press the Numpad0 key on the numeric keypad, or even if they type Alt + Numpad4 , Numpad8 to type the character by entering its character code on the numeric keypad. In the above case, the accelerator was defined as the character ^M , which is shorthand for Ctrl+M, or character co...

As we close out 2025, we want to pause and reflect on what has been an extraordinary year for the SharePoint Framework (SPFx). We continue to be humbled by the momentum of the platform and the scale of impact it enables across Microsoft 365. Tens of millions of users around the world rely every day on SPFx-based solutions - solutions imagined, built, and refined by our incredible ecosystem of customers and partners. Since launching this monthly blog series in September 2025, our goal has been to provide greater transparency, clearer communication, and a consistent view into the SPFx roadmap . With this December 2025 update, we’re taking the opportunity not only to share the latest roadmap perspective, but also to look back on the collective progress we’ve made throughout the year. From continued platform stability and performance improvements to deeper engagement with the community, 2025 has reinforced SPFx as a foundational pillar of Microsoft 365 extensibility. None of this prog...

Amazon CloudWatch now allows customers to automatically configure telemetry across their AWS Organization for six critical AWS services: AWS CloudTrail Management Events, AWS CloudTrail Data Events, Amazon Route 53 Resource Query Logs, Amazon EKS Control Plane logs, Network Load Balancer access logs, and AWS WAF WebACL Logs. This enhancement to CloudWatch enablement rules helps organizations maintain consistent monitoring and security practices at scale. With today's launch, customers can create organization-wide enablement rules that automatically configure logging for both existing and new resources across these services. For example, security teams can create rules to automatically configure CloudTrail Management Events and CloudTrail Data Events for all accounts in their organization, ensuring comprehensive audit trails. Operations teams can enforce consistent logging practices by automatically enabling EKS Control Plane logs across clusters with specific resource tags. Enabl...

Today, Amazon GameLift Streams launched two new capabilities to optimize performance and cost: Gen6 stream classes and enhanced autoscaling with warm buffer. The new Gen6 stream classes provide a wider range of price performance options, while autoscaling enables customers to dynamically manage capacity scaling.  The seven new Gen6 stream classes available today are based on EC2 G6 instances powered by NVIDIA L4 Tensor Core GPUs, which provide up to 2x higher performance over Gen4 stream classes. The  pro  and  ultra  stream class deliver improved performance for graphics-intensive AAA games, while the  medium  and  small  stream class offer cost-efficient options for casual games. The gen6n_small  stream class is available at $0.16/hour in us-east-2.  The enhanced autoscaling capabilities provide automatic capacity management that scales provisioned capacity dynamically with demand, helping customers optimize utilizat...

AWS Wickr now provides a suite of admin APIs that empower administrators to programmatically manage secure communication networks at scale. These APIs enable you to automate critical administrative workflows including user lifecycle management, network configuration, and security group administration. With user lifecycle management APIs, you can automatically create users and assign security groups when employees join, or deactivate accounts when they leave. Network configuration APIs allow you to quickly create or delete networks on demand as your organization scales or restructures, and push standardized retention and federation policies across departments. Security group administration APIs enable automatic user placement based on directory attributes such as job function or clearance level. By connecting Wickr administration directly into your identity management systems, policy management frameworks, and automation pipelines, you can now manage secure communications infrastructure...

Many languages¹ that have exceptions also have a finally clause, so you can write try { ⟦ stuff ⟧ } finally { always(); } A quick checks shows that this control structure exists in Java, C#, Python, JavaScript, but not C++. C++ says, "We have try ...finally at home." In C++, the way to get a block of code to execute when control leaves a block is to put it in a destructor, because destructors run when control leaves a block. This is the trick used by the Windows Implementation Library's wil:: scope_exit function: The lambda you provide is placed inside a custom object whose destructor runs the lambda. auto ensure_cleanup = wil::scope_exit([&] { always(); }); ⟦ stuff ⟧ Although the principle is the same, there are some quirks in how each language treats the case where the finally or destructor itself throws an exception. If control leaves the block without an exception, then any uncaught exception that occurs in the finally block or the destructor i...

Delivering high-quality software requires clarity, speed, and collaboration. That’s why we introduced the New Test Run Hub in Azure Test Plans. A modern, streamlined experience designed to make test execution and analysis fast and intuitive. And we’re excited to announce that this experience is moving to General Availability (GA) for the Azure DevOps Services throughout January 2026. Why the New Test Run Hub? The new hub centralizes test execution for both manual and automated runs, giving teams: Real-Time Visibility: Instantly monitor test progress and quality trends to catch regressions before they impact your release. Comprehensive Analytics: Dive into historical data with built-in dashboards that break down results by outcome, priority, configuration, and failure type. Effortless Management: Use powerful filters such as timeline, run type, pipeline, and more, to find exactly what you need. Customize your view with persistent search and column visibility settings. Rich Coll...

Amazon Aurora PostgreSQL-Compatible Edition has added support for PostgreSQL versions 17.7, 16.11, 15.15, 14.20 and 13.23. The update includes the PostgreSQL community's product improvements and bug fixes, and also includes Aurora-specific enhancements. The release includes improvements to Blue Green deployments, resulting in reduced switchover times by limiting new commits on the primary instance. There are also enhancements to Query Plan Management (QPM), and improved instance recovery times by optimizing commit log loading during database recovery. To use the new versions, create a new Aurora PostgreSQL-compatible database with just a few clicks in the Amazon RDS Management Console. You can also upgrade your existing database. Please review the Aurora documentation to learn more about upgrading. Refer to the Aurora version policy to help you to decide how often to upgrade and how to plan your upgrade process. These releases are available in all commercial AWS Regions and ...

Amazon WorkSpaces Applications now offers images powered by Microsoft Windows Server 2025, enabling customers to launch streaming instances with the latest features and enhancements from Microsoft’s newest server operating system. This update ensures your application streaming environment benefits from improved security, performance, and modern capabilities. With Windows Server 2025 support, you can deliver the Microsoft Windows 11 desktop experience to your end users, giving you greater flexibility in choosing the right operating system for your specific application and desktop streaming needs. Whether you're running business-critical applications or providing remote access to specialized software, you now have expanded options to align your infrastructure decisions with your unique workload requirements and organizational standards. You can select from AWS-provided public images or create custom images tailored to your requirements using Image Builder. Support for Microsoft W...

Security vulnerabilities don't fix themselves. Someone needs to track them, prioritize them, and actually ship the fix. If you've ever tried to manage security alerts alongside your regular sprint work, though, you know the friction: you're looking at an alert in one tab, switching to your backlog in another, trying to remember which vulnerability you were supposed to file a bug for. We shipped work item linking for GitHub Advanced Security for Azure DevOps alerts to fix this. It's now generally available and it does exactly what it sounds like: you can link work items in Boards directly to security alerts. Note that this only works for Advanced Security alerts in Azure DevOps. The problem we see Security alerts live in the Advanced Security hub while sprint planning happens in Boards. Teams end up with lost context (which alerts have owners?) and visibility gaps (is anyone actually working on this vulnerability?). When your security team asks "is someone fixin...

Say you have a shortcut file and you want to see what its target is. The natural thing to do is to call IShellLink:: GetPath , which returns S_FALSE if the shortcut does not target a path, or returns S_OK and a path to the target. But for some shortcuts, it returns S_OK but produces a weird path to a file inside the C:\Windows\ Installer directory that is definitely not the shortcut target. So what is it? You're seeing a backward compatibility consequence of shortcuts to MSI applications. The Microsoft Windows Installer (MSI) was introduced in Windows 2000. It supports various types of installation functionality, but the interesting one here is install-on-demand. A shortcut to an MSI-installed application is not a shortcut to an executable, but rather it is a shortcut that encodes the MSI information to identify an MSI application. When the user double-clicks the shortcut, the shortcut file asks MSI to get the application ready to run, which involves installing the applicatio...

TL;DR Move from prototype to production in hours, not weeks: The new Microsoft Agent Framework and Hosted Agents let you build, test, and deploy multi-agent AI systems with enterprise-grade security—no Kubernetes or container headaches. Orchestrate any model, anywhere: Model Router and BYO Model Gateway let you mix and match thousands of models (including Claude, GPT, and your own) with unified governance and compliance—no code changes required. Ship agents to Teams and M365 with one click: New low-code/no-code tools, templates, and deployment channels make it easy to launch and scale AI agents for your users. Build smarter, more reliable workflows: Multi-agent orchestration, persistent memory, and deep Microsoft 365 integration enable robust, context-aware solutions for complex enterprise scenarios. Fine-tune and innovate faster: Redesigned UI, support for reinforcement fine-tuning (RFT) on GPT-5, and parity for non-OpenAI models like Mistral accelerate custom model develop...