Self-Scaling LLM Infrastructure Powered by Rancher by ademicev0

Self-Scaling LLM Infrastructure Powered by Rancher

---

Description

The Problem

Running LLMs can get expensive and complex pretty quickly.

Today there are typically two choices:

1. Use cloud APIs like OpenAI or Anthropic. Easy to start with, but costs add up at scale.
2. Self-host everything - set up Kubernetes, figure out GPU scheduling, handle scaling, manage model serving... it's a lot of work.

What if there was a middle ground?

What if infrastructure scaled itself instead of making you scale it?

Can we use existing Rancher capabilities like CAPI, autoscaling, and GitOps to make this simpler instead of building everything from scratch?

Project Repository: github.com/alexander-demicev/llmserverless

---

What This Project Does

A key feature is hybrid deployment: requests can be routed based on complexity or privacy needs. Simple or low-sensitivity queries can use public APIs (like OpenAI), while complex or private requests are handled in-house on local infrastructure. This flexibility allows balancing cost, privacy, and performance - using cloud for routine tasks and on-premises resources for sensitive or demanding workloads.

A complete, self-scaling LLM infrastructure that:

• Scales to zero when idle (no idle costs)
• Scales up automatically when requests come in
• Adds more nodes when needed, removes them when demand drops
• Runs on any infrastructure - laptop, bare metal, or cloud

Think of it as "serverless for LLMs" - focus on building, the infrastructure handles itself.

How It Works

A combination of open source tools working together:

Flow:

• Users interact with OpenWebUI (chat interface)
• Requests go to LiteLLM Gateway
• LiteLLM routes requests to:
  • Ollama (Knative) for local model inference (auto-scales pods)
  • Or cloud APIs for fallback

---

The Agentic Rancher Experiment: Do Androids Dream of Electric Cattle? by moio

Rancher is a beast of a codebase. Let's investigate if the new 2025 generation of GitHub Autonomous Coding Agents and Copilot Workspaces can actually tame it.

---

The Plan

Create a sandbox GitHub Organization, clone in key Rancher repositories, and let the AI loose to see if it can handle real-world enterprise OSS maintenance - or if it just hallucinates new breeds of Kubernetes resources!

Specifically, throw "Agentic Coders" some typical tasks in a complex, long-lived open-source project, such as:

❥ The Grunt Work: generate missing GoDocs, unit tests, and refactorings. Rebase PRs.

❥ The Complex Stuff: fix actual (historical) bugs and feature requests to see if they can traverse the complexity without (too much) human hand-holding.

❥ Hunting Down Gaps: find areas lacking in docs, areas of improvement in code, dependency bumps, and so on.

If time allows, also experiment with Model Context Protocol (MCP) to give agents context on our specific build pipelines and CI/CD logs.

Why?

We know AI can write "Hello World." and also moderately complex programs from a green field. But can it rebase a 3-month-old PR with conflicts in rancher/rancher? I want to find the breaking point of current AI agents to determine if and how they can help us to reduce our technical debt, work faster and better. At the same time, find out about pitfalls and shortcomings.

The CONCLUSION!!!

A State of the Union document was compiled to summarize lessons learned this week. For more gory details, just read on the diary below!

---

Extended private brain - RAG my own scripts and data into offline LLM AI by tjyrinki_suse

Description

For purely studying purposes, I'd like to find out if I could teach an LLM some of my own accumulated knowledge, to use it as a sort of extended brain.

I might use qwen3-coder or something similar as a starting point.

Everything would be done 100% offline without network available to the container, since I prefer to see when network is needed, and make it so it's never needed (other than initial downloads).

Goals

1. Learn something about RAG, LLM, AI.
2. Find out if everything works offline as intended.
3. As an end result have a new way to access my own existing know-how, but so that I can query the wisdom in them.
4. Be flexible to pivot in any direction, as long as there are new things learned.

Resources

To be found on the fly.

Timeline

Day 1 (of 4)

• Tried out a RAG demo, expanded on feeding it my own data
• Experimented with qwen3-coder to add a persistent chat functionality, and keeping vectors in a pickle file
• Optimizations to keep everything within context window
• Learn and add a bit of PyTest

Day 2

• More experimenting and more data
• Study ChromaDB
• Add a Web UI that works from another computer even though the container sees network is down

Day 3

• The above RAG is working well enough for demonstration purposes.
• Pivot to trying out OpenCode, configuring local Ollama qwen3-coder there, to analyze the RAG demo.
• Figured out how to configure Ollama template to be usable under OpenCode. OpenCode locally is super slow to just running qwen3-coder alone.

Day 4 (final day)

• Battle with OpenCode that was both slow and kept on piling up broken things.
• Call it success as after all the agentic AI was working locally.
• Clean up the mess left behind a bit.

Blog Post

Summarized the findings at blog post.

---

Exploring Modern AI Trends and Kubernetes-Based AI Infrastructure by jluo

Description

Build a solid understanding of the current landscape of Artificial Intelligence and how modern cloud-native technologies—especially Kubernetes—support AI workloads.

Goals

Use Gemini Learning Mode to guide the exploration, surface relevant concepts, and structure the learning journey:

• Gain insight into the latest AI trends, tools, and architectural concepts.
• Understand how Kubernetes and related cloud-native technologies are used in the AI ecosystem (model training, deployment, orchestration, MLOps).

Resources

• Red Hat AI Topic Articles

  • https://www.redhat.com/en/topics/ai

• Kubeflow Documentation

  • https://www.kubeflow.org/docs/

• Q4 2025 CNCF Technology Landscape Radar report:

  • https://www.cncf.io/announcements/2025/11/11/cncf-and-slashdata-report-finds-leading-ai-tools-gaining-adoption-in-cloud-native-ecosystems/
  • https://www.cncf.io/wp-content/uploads/2025/11/cncfreporttechradar_111025a.pdf

• Agent-to-Agent (A2A) Protocol

  • https://developers.googleblog.com/en/a2a-a-new-era-of-agent-interoperability/

---

SUSE Edge Image Builder MCP by eminguez

Description

Based on my other hackweek project, SUSE Edge Image Builder's Json Schema I would like to build also a MCP to be able to generate EIB config files the AI way.

Realistically I don't think I'll be able to have something consumable at the end of this hackweek but at least I would like to start exploring MCPs, the difference between an API and MCP, etc.

Goals

• Familiarize myself with MCPs
• Unrealistic: Have an MCP that can generate an EIB config file

Resources

• https://hackweek.opensuse.org/25/projects/suse-edge-image-builder-json-schema
• Anything else :)

Result

https://github.com/e-minguez/eib-mcp

I've extensively used antigravity and its agent mode to code this. This heavily uses https://hackweek.opensuse.org/25/projects/suse-edge-image-builder-json-schema for the MCP to be built.

I've ended up learning a lot of things about "prompting", json schemas in general, some golang, MCPs and AI in general :)

Example:

Generate an Edge Image Builder configuration for an ISO image based on slmicro-6.2.iso, targeting x86_64 architecture. The output name should be 'my-edge-image' and it should install to /dev/sda. It should deploy a 3 nodes kubernetes cluster with nodes names "node1", "node2" and "node3" as: * hostname: node1, IP: 1.1.1.1, role: initializer * hostname: node2, IP: 1.1.1.2, role: agent * hostname: node3, IP: 1.1.1.3, role: agent The kubernetes version should be k3s 1.33.4-k3s1 and it should deploy a cert-manager helm chart (the latest one available according to https://cert-manager.io/docs/installation/helm/). It should create a user called "suse" with password "suse" and set ntp to "foo.ntp.org". The VIP address for the API should be 1.2.3.4

Generates:

``` apiVersion: "1.0" image: arch: x86_64 baseImage: slmicro-6.2.iso imageType: iso outputImageName: my-edge-image kubernetes: helm: charts: - name: cert-manager repositoryName: jetstack

---

openQA log viewer by mpagot

Description

*** Warning: Are You at Risk for VOMIT? ***

Do you find yourself staring at a screen, your eyes glossing over as thousands of lines of text scroll by? Do you feel a wave of text-based nausea when someone asks you to "just check the logs"?

You may be suffering from VOMIT (Verbose Output Mental Irritation Toxicity).

This dangerous, work-induced ailment is triggered by exposure to an overwhelming quantity of log data, especially from parallel systems. The human brain, not designed to mentally process 12 simultaneous autoinst-log.txt files, enters a state of toxic shock. It rejects the "Verbose Output," making it impossible to find the one critical error line buried in a 50,000-line sea of "INFO: doing a thing."

Before you're forced to rm -rf /var/log in a fit of desperation, we present the digital antacid.

No panic: we have The openQA Log Visualizer

This is the UI antidote for handling toxic log environments. It bravely dives into the chaotic, multi-machine mess of your openQA test runs, finds all the related, verbose logs, and force-feeds them into a parser.

image

Goals

Work on the existing POC openqa-log-visualizer about few specific tasks:

• add support for more type of logs
• extend the configuration file syntax beyond the actual one
• work on log parsing performance

Find some beta-tester and collect feedback and ideas about features

If time allow for it evaluate other UI frameworks and solutions (something more simple to distribute and run, maybe more low level to gain in performance).

Resources

openqa-log-visualizer

---

openQA tests needles elaboration using AI image recognition by mdati

Description

In the openQA test framework, to identify the status of a target SUT image, a screenshots of GUI or CLI-terminal images, the needles framework scans the many pictures in its repository, having associated a given set of tags (strings), selecting specific smaller parts of each available image. For the needles management actually we need to keep stored many screenshots, variants of GUI and CLI-terminal images, eachone accompanied by a dedicated set of data references (json).

A smarter framework, using image recognition based on AI or other image elaborations tools, nowadays widely available, could improve the matching process and hopefully reduce time and errors, during the images verification and detection process.

Goals

Main scope of this idea is to match a "graphical" image of the console or GUI status of a running openQA test, an image of a shell console or application-GUI screenshot, using less time and resources and with less errors in data preparation and use, than the actual openQA needles framework; that is:

• having a given SUT (system under test) GUI or CLI-terminal screenshot, with a local distribution of pixels or text commands related to a running test status,
• we want to identify a desired target, e.g. a screen image status or data/commands context,
  • based on AI/ML-pretrained archives containing object or other proper elaboration tools,
  • possibly able to identify also object not present in the archive, i.e. by means of AI/ML mechanisms.
• the matching result should be then adapted to continue working in the openQA test, likewise and in place of the same result that would have been produced by the original openQA needles framework.
• We expect an improvement of the matching-time(less time), reliability of the expected result(less error) and simplification of archive maintenance in adding/removing objects(smaller DB and less actions).

Hackweek POC:

Main steps

• Phase 1 - Plan
  • study the available tools
  • prepare a plan for the process to build
• Phase 2 - Implement
  • write and build a draft application
• Phase 3 - Data
  • prepare the data archive from a subset of needles
  • initialize/pre-train the base archive
  • select a screenshot from the subset, removing/changing some part
• Phase 4 - Test
  • run the POC application
  • expect the image type is identified in a good %.

Resources

First step of this project is quite identification of useful resources for the scope; some possibilities are:

• SUSE AI and other ML tools (i.e. Tensorflow)
• Tools able to manage images
• RPA test tools (like i.e. Robot framework)
  
• other.

Project references

• Repository: openqa-needles-AI-driven

---

MCP Perl SDK by kraih

Description

We've been using the MCP Perl SDK to connect openQA with AI. And while the basics are working pretty well, the SDK is not fully spec compliant yet. So let's change that!

Goals

• Support for Resources
• All response types (Audio, Resource Links, Embedded Resources...)
• Tool/Prompt/Resource update notifications
• Dynamic Tool/Prompt/Resource lists
• New authentication mechanisms

Resources

• https://github.com/mojolicious/mojo-mcp
• https://modelcontextprotocol.io/specification/2025-06-18

---

Bring up Agama based tests for openSUSE Tumbleweed by szarate

Description

Agama has been around for some time already, and we have some tests for it on Tumbleweed however they are only on the development job group and are too few to be helpful in assessing the quality of a build

This project aims at enabling and creating new testsuites for the agama flavor, using the already existsing DVD and NET flavors as starting points

Goals

• Introduce tests based on the Agama flavor in the main Tumbleweed job group
• Create Tumbleweed yaml schedules for agama installer and its own jsonette profile (The one being used now are reused from leap)
• Fan out tests that have long runtimes (i.e tackle this ticket)
• Reduce redundancy in tests

Resources

• Tumbleweed development job group:
• Tumbleweed main job group in git
• osado test repository:

---

Testing and adding GNU/Linux distributions on Uyuni by juliogonzalezgil

Join the Gitter channel! https://gitter.im/uyuni-project/hackweek

Uyuni is a configuration and infrastructure management tool that saves you time and headaches when you have to manage and update tens, hundreds or even thousands of machines. It also manages configuration, can run audits, build image containers, monitor and much more!

Currently there are a few distributions that are completely untested on Uyuni or SUSE Manager (AFAIK) or just not tested since a long time, and could be interesting knowing how hard would be working with them and, if possible, fix whatever is broken.

For newcomers, the easiest distributions are those based on DEB or RPM packages. Distributions with other package formats are doable, but will require adapting the Python and Java code to be able to sync and analyze such packages (and if salt does not support those packages, it will need changes as well). So if you want a distribution with other packages, make sure you are comfortable handling such changes.

No developer experience? No worries! We had non-developers contributors in the past, and we are ready to help as long as you are willing to learn. If you don't want to code at all, you can also help us preparing the documentation after someone else has the initial code ready, or you could also help with testing :-)

The idea is testing Salt (including bootstrapping with bootstrap script) and Salt-ssh clients

To consider that a distribution has basic support, we should cover at least (points 3-6 are to be tested for both salt minions and salt ssh minions):

1. Reposync (this will require using spacewalk-common-channels and adding channels to the .ini file)
2. Onboarding (salt minion from UI, salt minion from bootstrap scritp, and salt-ssh minion) (this will probably require adding OS to the bootstrap repository creator)
3. Package management (install, remove, update...)
4. Patching
5. Applying any basic salt state (including a formula)
6. Salt remote commands
7. Bonus point: Java part for product identification, and monitoring enablement
8. Bonus point: sumaform enablement (https://github.com/uyuni-project/sumaform)
9. Bonus point: Documentation (https://github.com/uyuni-project/uyuni-docs)
10. Bonus point: testsuite enablement (https://github.com/uyuni-project/uyuni/tree/master/testsuite)

If something is breaking: we can try to fix it, but the main idea is research how supported it is right now. Beyond that it's up to each project member how much to hack :-)

• If you don't have knowledge about some of the steps: ask the team
• If you still don't know what to do: switch to another distribution and keep testing.

This card is for EVERYONE, not just developers. Seriously! We had people from other teams helping that were not developers, and added support for Debian and new SUSE Linux Enterprise and openSUSE Leap versions :-)

In progress/done for Hack Week 25

Guide

We started writin a Guide: Adding a new client GNU Linux distribution to Uyuni at https://github.com/uyuni-project/uyuni/wiki/Guide:-Adding-a-new-client-GNU-Linux-distribution-to-Uyuni, to make things easier for everyone, specially those not too familiar wht Uyuni or not technical.

openSUSE Leap 16.0

The distribution will all love!

https://en.opensuse.org/openSUSE:Roadmap#DRAFTScheduleforLeap16.0

Curent Status We started last year, it's complete now for Hack Week 25! :-D

• [W] Reposync (this will require using spacewalk-common-channels and adding channels to the .ini file) NOTE: Done, client tools for SLMicro6 are using as those for SLE16.0/openSUSE Leap 16.0 are not available yet
• [W] Onboarding (salt minion from UI, salt minion from bootstrap scritp, and salt-ssh minion) (this will probably require adding OS to the bootstrap repository creator)
• [W] Package management (install, remove, update...). Works, even reboot requirement detection

---

Multimachine on-prem test with opentofu, ansible and Robot Framework by apappas

Description

A long time ago I explored using the Robot Framework for testing. A big deficiency over our openQA setup is that bringing up and configuring the connection to a test machine is out of scope.

Nowadays we have a way¹ to deploy SUTs outside openqa, but we only use if for cloud tests in conjuction with openqa. Using knowledge gained from that project I am going to try to create a test scenario that replicates an openqa test but this time including the deployment and setup of the SUT.

Goals

Create a simple multimachine test scenario with the support server and SUT all created by the robot framework.

Resources

1. https://github.com/SUSE/qe-sap-deployment
2. terraform-libvirt-provider

---

openQA tests needles elaboration using AI image recognition by mdati

Description

In the openQA test framework, to identify the status of a target SUT image, a screenshots of GUI or CLI-terminal images, the needles framework scans the many pictures in its repository, having associated a given set of tags (strings), selecting specific smaller parts of each available image. For the needles management actually we need to keep stored many screenshots, variants of GUI and CLI-terminal images, eachone accompanied by a dedicated set of data references (json).

A smarter framework, using image recognition based on AI or other image elaborations tools, nowadays widely available, could improve the matching process and hopefully reduce time and errors, during the images verification and detection process.

Goals

Main scope of this idea is to match a "graphical" image of the console or GUI status of a running openQA test, an image of a shell console or application-GUI screenshot, using less time and resources and with less errors in data preparation and use, than the actual openQA needles framework; that is:

• having a given SUT (system under test) GUI or CLI-terminal screenshot, with a local distribution of pixels or text commands related to a running test status,
• we want to identify a desired target, e.g. a screen image status or data/commands context,
  • based on AI/ML-pretrained archives containing object or other proper elaboration tools,
  • possibly able to identify also object not present in the archive, i.e. by means of AI/ML mechanisms.
• the matching result should be then adapted to continue working in the openQA test, likewise and in place of the same result that would have been produced by the original openQA needles framework.
• We expect an improvement of the matching-time(less time), reliability of the expected result(less error) and simplification of archive maintenance in adding/removing objects(smaller DB and less actions).

Hackweek POC:

Main steps

• Phase 1 - Plan
  • study the available tools
  • prepare a plan for the process to build
• Phase 2 - Implement
  • write and build a draft application
• Phase 3 - Data
  • prepare the data archive from a subset of needles
  • initialize/pre-train the base archive
  • select a screenshot from the subset, removing/changing some part
• Phase 4 - Test
  • run the POC application
  • expect the image type is identified in a good %.

Resources

First step of this project is quite identification of useful resources for the scope; some possibilities are:

• SUSE AI and other ML tools (i.e. Tensorflow)
• Tools able to manage images
• RPA test tools (like i.e. Robot framework)
  
• other.

Project references

• Repository: openqa-needles-AI-driven

---

Enhance git-sha-verify: A tool to checkout validated git hashes by gpathak

Description

git-sha-verify is a simple shell utility to verify and checkout trusted git commits signed using GPG key. This tool helps ensure that only authorized or validated commit hashes are checked out from a git repository, supporting better code integrity and security within the workflow.

Supports:

• Verifying commit authenticity signed using gpg key
  
• Checking out trusted commits
  

Ideal for teams and projects where the integrity of git history is crucial.

Goals

A minimal python code of the shell script exists as a pull request.

The goal of this hackweek is to:

• DONE: Add more unit tests
  • New and more tests can be added later
    
• Partially DONE: Make the python code modular
  
• DONE: Add code coverage if possible
  

Resources

• Link to GitHub Repository: https://github.com/openSUSE/git-sha-verify

---

Collection and organisation of information about Bulgarian schools by iivanov

Description

To achieve this it will be necessary:

• Collect/download raw data from various government and non-governmental organizations
• Clean up raw data and organise it in some kind database.
• Create tool to make queries easy.
• Or perhaps dump all data into AI and ask questions in natural language.

Goals

By selecting particular school information like this will be provided:

• School scores on national exams.
• School scores from the external evaluations exams.
• School town, municipality and region.
• Employment rate in a town or municipality.
• Average health of the population in the region.

Resources

Some of these are available only in bulgarian.

• https://danybon.com/klasazia
• https://nvoresults.com/index.html
• https://ri.mon.bg/active-institutions
• https://www.nsi.bg/nrnm/ekatte/archive

Results

• Information about all Bulgarian schools with their scores during recent years cleaned and organised into SQL tables
• Information about all Bulgarian villages, cities, municipalities and districts cleaned and organised into SQL tables
• Information about all Bulgarian villages and cities census since beginning of this century cleaned and organised into SQL tables.
• Information about all Bulgarian municipalities about religion, ethnicity cleaned and organised into SQL tables.
• Data successfully loaded to locally running Ollama with help to Vanna.AI
• Seems to be usable.

TODO

• Add more statistical information about municipalities and ....

Code and data

• Github

---

Bring to Cockpit + System Roles capabilities from YAST by miguelpc

Bring to Cockpit + System Roles features from YAST

Cockpit and System Roles have been added to SLES 16 There are several capabilities in YAST that are not yet present in Cockpit and System Roles We will follow the principle of "automate first, UI later" being System Roles the automation component and Cockpit the UI one.

Goals

The idea is to implement service configuration in System Roles and then add an UI to manage these in Cockpit. For some capabilities it will be required to have an specific Cockpit Module as they will interact with a reasource already configured.

Resources

A plan on capabilities missing and suggested implementation is available here: https://docs.google.com/spreadsheets/d/1ZhX-Ip9MKJNeKSYV3bSZG4Qc5giuY7XSV0U61Ecu9lo/edit

Linux System Roles:

• https://linux-system-roles.github.io/
• https://build.opensuse.org/package/show/openSUSE:Factory/ansible-linux-system-roles Package on sle16 ansible-linux-system-roles

First meeting Hackweek catchup

• Monday, December 1 · 11:00 – 12:00
• Time zone: Europe/Madrid
• Google Meet link: https://meet.google.com/rrc-kqch-hca

---

Update M2Crypto by mcepl

There are couple of projects I work on, which need my attention and putting them to shape:

• M2Crypto

Goal for this Hackweek

• Put M2Crypto into better shape (most issues closed, all pull requests processed)
• More fun to learn jujutsu
• Play more with Gemini, how much it help (or not).
• Perhaps, also (just slightly related), help to fix vis to work with LuaJIT, particularly to make vis-lspc working.

---

Song Search with CLAP by gcolangiuli

Description

Contrastive Language-Audio Pretraining (CLAP) is an open-source library that enables the training of a neural network on both Audio and Text descriptions, making it possible to search for Audio using a Text input. Several pre-trained models for song search are already available on huggingface

Goals

Evaluate how CLAP can be used for song searching and determine which types of queries yield the best results by developing a Minimum Viable Product (MVP) in Python. Based on the results of this MVP, future steps could include:

• Music Tagging;
• Free text search;
• Integration with an LLM (for example, with MCP or the OpenAI API) for music suggestions based on your own library.

The code for this project will be entirely written using AI to better explore and demonstrate AI capabilities.

Result

In this MVP we implemented:

• Async Song Analysis with Clap model
• Free Text Search of the songs
• Similar song search based on vector representation
• Containerised version with web interface

We also documented what went well and what can be improved in the use of AI.

You can have a look at the result here:

• Result Presentation
• hw25-song-search Github repo

Future implementation can be related to performance improvement and stability of the analysis.

References

• CLAP: The main model being researched;
• huggingface: Pre-trained models for CLAP;
• Free Music Archive: Creative Commons songs that can be used for testing;

---