How to Run GitHub Actions Runner Controller on AKS Automatic: Real-World Lessons

·1 views

On 17 September 2026, the AKS engineering blog published a guide by Steve Griffith detailing how to deploy GitHub Actions Runner Controller (ARC) on an AKS Automatic cluster. A developer followed the guide end-to-end on a fresh Azure subscription and documented the exact commands, outputs, and deviations from the original instructions. Three notable issues arose: an outdated Azure CLI caused Python tracebacks, insufficient vCPU quota blocked cluster creation, and the cluster took 33 minutes to provision rather than the expected ten. Quota requests for newer AMD v6 VM families were approved instantly via CLI, while requests for DSv5 families required contacting support. The accompanying demo repository includes one make target per step, troubleshooting notes, and a validation workflow that can be forked and pointed at any GitHub repository.

Read the full story at DEV Community

This is an AI-generated summary. ShortSingh links to the original source for the complete article.

Discussion (0)

Why AI-Generated Code Needs Intent-Level Tests, Not Just Green Checks

Modern large language models can generate code that compiles cleanly and passes standard unit tests while still violating critical business rules, such as security constraints or financial precision requirements. Unlike human developers, LLMs optimize for statistical pattern matching rather than logical intent, making them prone to producing confidently broken code. A three-layer testing hierarchy is proposed — covering syntax, behavior, and domain invariants — with most AI-generated code passing only the first layer. Intent-level validation strategies such as property-based testing, state-machine verification, and anti-pattern suites are recommended to catch logical and business-rule failures. Integrating these checks into CI/CD pipelines is presented as a necessary step for teams relying on AI-assisted development.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Google Opens Googlebook Preorders at $899, Launching Gemini-Powered Laptop Category

Google has begun accepting preorders for Googlebook, a new laptop category built around its Gemini AI and tight Android integration, starting at $899. Devices are set to go on sale on October 4, 2025 in the United States and October 5 in other markets. The hardware will be manufactured by Acer, ASUS, Dell, HP and Lenovo, while Google controls the software layer through its new Googlebook OS. Buyers will receive a bundle that includes 12 months of Google AI Pro, 5 TB of cloud storage, Gemini Advanced, and limited-period YouTube Premium and Photoshop benefits. Google has committed to software updates for up to 10 years, underscoring its intent to position Googlebook as a long-term AI-driven device platform rather than a one-time hardware release.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

IBM Guardium Data Protection 12.2 Patches 18 CVEs Including 10 Critical Flaws

IBM Guardium Data Protection version 12.2 is affected by a group of over 20 vulnerabilities disclosed together in September 2026, comprising 18 CVEs with 10 rated Critical and 7 rated High. Among them, CVE-2026-81657 carries a severity score of 9.8 and involves an insecure deserialization flaw in the Change Audit System listener on TCP port 16017, allowing unauthenticated remote attackers to potentially execute arbitrary code. The vulnerabilities affect core data protection components, load balancers, and administrative web interfaces across the Guardium product family. IBM has released patches and is urging customers to apply them promptly, recommending a single consolidated update pass given the batch nature of the disclosures. Until patching is complete, administrators are advised to restrict access to TCP port 16017, limit management interface exposure to trusted hosts, and review authentication logs for suspicious activity.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Django, Flask or FastAPI: Which Python Web Framework to Pick in 2026

Search interest in Python web frameworks has surged 190% in the UK over the past three months, driven by Python's growing dominance in AI integration and backend development. The three leading frameworks — Django, Flask, and FastAPI — share the same language but serve distinct use cases and developer needs. Django is recommended for full-stack applications requiring complex data models, built-in admin tools, and strong security defaults, making it well-suited for regulated industries. Flask offers a lightweight, flexible alternative for microservices where developers prefer to assemble their own stack with minimal framework constraints. FastAPI is highlighted as the strongest default choice for new Python API projects in 2026, particularly those involving AI workloads or high concurrency requirements.

0 comments Read more at DEV Community