<h2>Must know</h2>
<ul>
<li><strong>Kubernetes ingress-nginx is getting archived</strong>: The Kubernetes ingress-nginx project is being archived in early 2026, meaning no more active maintenance or security updates. This affects roughly half of all cloud-native environments, forcing users to migrate to alternatives like Cilium Ingress or Gateway API. The situation highlights the ongoing challenges of open-source sustainability and maintainer burnout, as reported by DEVCLASS and CNCF. <a href="https://app.daily.dev/feed-by-ids?id=PptBbhnAu&amp;id=wK9Dxbori" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Upwind security raises big bucks for runtime cloud security</strong>: Upwind Security just snagged $250 million in Series B funding, valuing them at $1.5 billion, to expand their ‘runtime’ cloud security platform. They focus on an ‘inside-out’ approach, analyzing internal signals like network requests and API traffic to prioritize real-time threats, a contrast to traditional agentless scanning. This funding will fuel their AI security capabilities and misconfiguration prevention efforts, as detailed by TechCrunch. <a href="https://app.daily.dev/feed-by-ids?id=LGgCBPA2D&amp;id=Upwind+Security+Secures+%24250M+Series+B+Funding" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Google Cloud Run now supports NVIDIA RTX 6000 Pro GPUs</strong>: Google Cloud Run is now offering NVIDIA RTX PRO 6000 Blackwell GPUs in preview, enabling serverless deployment of large AI models without the hassle of cluster management. These instances can scale from zero to ready in under 5 seconds, supporting models up to 70B+ parameters. This is a significant step for AI workloads requiring high-performance inference at the edge, as announced by Google Cloud. <a href="https://app.daily.dev/posts/Vh4SzEb3w" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>New SQLite-compatible edge database from Bunny</strong>: Bunny Database, built on libSQL, is a new SQLite-compatible managed database service offering global deployment, low-latency reads, and pay-as-you-go pricing, with databases spinning down when idle to minimize costs. <a href="https://app.daily.dev/posts/bbnTGzjvp" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>ClickHouse Cloud slashes costs and boosts analytics for Buildkite</strong>: Buildkite migrated its Test Engine analytics to ClickHouse Cloud, retiring DynamoDB and reducing Aurora/Flink usage, resulting in five-figure monthly cost savings while quadrupling ingestion to 12 billion events per month. <a href="https://app.daily.dev/posts/JAWEznKml" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Google Cloud Spanner evolves into an AI context hub</strong>: Google Cloud Spanner has transformed into an intelligent context hub for AI applications, offering multi-model support, AI-powered operations, unified analytics, and enhanced enterprise safety features. <a href="https://app.daily.dev/posts/TRihFu0x6" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Managed connection pooling for Google AlloyDB for PostgreSQL</strong>: Google Cloud has launched managed connection pooling for AlloyDB for PostgreSQL, delivering 3x more client connections and up to 5x higher transactional throughput by automating PgBouncer-like functionality. <a href="https://app.daily.dev/posts/3Jt7lYzbz" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Faster pipelines with Apache Arrow and ADBC</strong>: Benchmarking shows Apache Arrow with ADBC provides a 3.7x speedup over SQLAlchemy for data pipelines by streaming data in columnar format directly to the destination, eliminating Python object serialization overhead. <a href="https://app.daily.dev/posts/ntB759wMg" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Kthena orchestrates LLM inference on Kubernetes</strong>: Kthena is a new open-source sub-project of Volcano designed for LLM inference orchestration on Kubernetes, addressing challenges like low GPU utilization and multi-model management through intelligent routing and scheduling. <a href="https://app.daily.dev/posts/HONeqoore" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Varnish Software launches sovereign CDN for Europe</strong>: Varnish Software launched Varnish CDN, a Europe-hosted content delivery network that keeps all traffic, logs, and metadata within European jurisdiction, addressing GDPR and NIS2 compliance requirements. <a href="https://app.daily.dev/posts/Hd4eStguc" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>AI-coded tools: because humans are overrated</strong>: GitHuman, a new local-first tool for reviewing AI-generated code, was apparently <em>entirely</em> written by Claude Code. And get this, every commit was human-reviewed using GitHuman itself! Because nothing says ‘trustworthy’ like an AI reviewing its own AI-generated code, right? It’s almost like they’re trying to cut out the middleman… which is us, the developers, as reported by the tool’s promotional material. <a href="https://app.daily.dev/posts/M78k9F0lW" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Discovery Insure’s AI driving instructor: because your car needs a nagging parent too</strong>: Discovery Insure launched an AI-powered Advanced Driving Instructor that uses telematics and generative AI to create ‘personalized driving goals and rewards.’ So now, your insurance company’s AI will judge your driving, set your goals, and give you ‘Discovery Miles’ for good behavior. Because who needs actual safe driving habits when you can gamify your commute with AI, as their very unbiased press release suggests? <a href="https://app.daily.dev/posts/LsXY4pcoJ" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Astrological CPU scheduler: because performance is written in the stars</strong>: Someone actually built a Linux kernel CPU scheduler that makes real scheduling decisions based on astrological principles. It uses planetary positions, zodiac signs, and lunar phases to determine task priorities. While it’s ‘explicitly for educational and entertainment purposes only,’ it’s a stark reminder that if you give engineers enough rope, they’ll build a CPU scheduler based on horoscopes. Because why rely on boring old algorithms when you can consult the cosmos for optimal task management? <a href="https://app.daily.dev/posts/4O3zUvO4w" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
