What's in the Toolbox?

Key Concepts

• Unovz: An open-source, modular data visualization framework built on D3.js, supporting multiple JavaScript frameworks (React, Angular, Svelte, etc.) and powering charting in F5 Distributed Cloud and potentially Big-IP.
• AI-Assisted Automation: Leveraging Large Language Models (LLMs) like Claude AI to automate network documentation, configuration, and analysis, but requiring careful oversight and iterative refinement.
• Home Lab Value: The importance of home labs for learning, experimentation, and skill development, advocating for an incremental approach to building them.
• AppWorld 2026: An F5 conference (March 10-12th in Las Vegas) featuring lab experiences, keynotes, breakout sessions, and a TCL “eye roll” contest.

Community Updates & Event Promotion

Jason Rom announced AppWorld 2026, taking place March 10-12th in Las Vegas at the Fountain Blue. The conference will feature lab experiences, keynotes, breakout sessions, and a TCL “eye roll” contest. Speakers throughout the stream emphasized the value of home labs and encouraged participation in AppWorld, potentially through contest entries utilizing tools like Unovz and I rule.

Unovz: A Modular Data Visualization Framework

Cho, from the F5 UI foundational team, introduced Unovz, an open-source data visualization framework. Unlike some frameworks that bundle all chart elements, Unovz is modular, requiring explicit addition of components like axes and tooltips. This approach allows for lightweight and highly customizable visualizations. Built on top of D3.js, Unovz leverages its scaling mechanisms (linear, ordinal) and supports multiple JavaScript frameworks including React, Angular, and Svelte through compilation to JavaScript. A live demonstration showcased the creation of a basic line chart using a CPU/GPU characteristics dataset, illustrating the process of importing data, defining scales, adding chart components, and incorporating tooltips and crosshairs. Cho emphasized that Unovz is freely available for use.

Home Lab Automation with AI: A Deep Dive

Jerry Lees detailed his home lab setup and his use of Claude AI to automate network documentation and analysis. His workflow involves scanning the network, generating a CSV of discovered devices and services, and using Claude to create documentation in both Markdown and MediaWiki formats. He highlighted the collaborative nature of using AI, emphasizing the need for existing knowledge and guiding the AI rather than expecting complete solutions. Techniques for managing context length and improving code quality include modularization and directive files. Jerry noted that Claude often requires “reigning in” due to a tendency to over-elaborate. He also pointed out the AI’s ability to suggest improvements he hadn’t considered.

Accelerating Cacti Configuration with AI

Building on the home lab automation theme, Jerry demonstrated using Claude to generate a Cacti configuration file directly from his existing lab documentation. By prompting Claude with the available information, he successfully created an import file that automatically generated all the necessary graphs within his Cacti instance, significantly reducing manual configuration effort. He confirmed Cacti’s backend is likely still based on Perl and highlighted AI’s value in assisting with languages the user isn’t proficient in, stating, “I can read Pearl. I'm not great at writing it.” He stressed the importance of not blindly accepting AI outputs, stating, “Don't do that. It’s dangerous.” and emphasized the need for careful oversight and iterative refinement. A key methodology involved providing Claude with a comprehensive summary of the codebase, allowing for more focused subsequent requests.

Technical Landscape

The discussion touched upon several technical terms and concepts, including: Unovz, D3.js, JavaScript Frameworks (React, Angular, Svelte), Kubernetes, Proxmox, Declarative Onboarding (AS3), MediaWiki, Claude AI, Cider Notation, Terraform, LXCs, UDCs, and QK Views. Hardware costs were also mentioned, with HP G3 mini PCs available for around $75, 128GB of DDR3 ECC memory for approximately $100, and two HP DL360 servers acquired for $150.

Conclusion

The stream showcased the power of open-source tools like Unovz and AI assistants like Claude in accelerating development and automation tasks. While AI offers significant potential for streamlining workflows, the speakers consistently emphasized the importance of human expertise, careful oversight, and iterative refinement. The value of home labs as a platform for learning and experimentation was also a recurring theme, with encouragement to start small and gradually expand capabilities. The upcoming AppWorld 2026 conference was presented as an opportunity to further explore these technologies and share experiences within the F5 community.