My Homelab in 3D: Inside the DGX Spark and the Homelab PC
An interactive 3D tour of my desk, an exploded DGX Spark and an i9 + RTX 5070 Ti homelab PC, plus how the two machines split the work and reach the internet.
Cloudutsuk
Hot topics
Kubernetes Agentic AI AI & LLMs DevSecOps Observability Terraform & IaC Cloud Security Docker HomeLab Interview Prep
Clusters, workloads, Helm and the interview questions that go with them.
View all 93 posts- My DGX Spark was holding 54 GB for models nobody was calling: scale-to-zero vLLM with KEDA
- Grafana stuck in Init:0/1: the initDashboards flag that never lets go
- Backstage SSO broke for 3 days because the portal started 21 seconds before Keycloak
- k3d's load balancer ran out of connections, and DNS took the whole cluster down
- eBPF Lab
- Cloud Container Services: AWS vs Azure vs Google Cloud
AI agents, tool calling and the Model Context Protocol (MCP).
View all 3 postsLLMs, local models, GenAI on Azure and AI coding tools.
View all 44 posts- My DGX Spark was holding 54 GB for models nobody was calling: scale-to-zero vLLM with KEDA
- The .claude Folder for Claude Code: Settings, Skills, Agents, Hooks and Rules Explained
- The .github Folder for GitHub Copilot: Instructions, Prompts, Agents, Skills and Hooks Explained
- Architecture of an AI Agent: How User, Agent, LLM and Tools Work Together
- Debugging a Stubborn TLS Handshake to get SonarQube MCP Server Back Online
- Claude Code - Tips and Tips
Security in the pipeline: SAST, SCA, secrets and supply chain.
View all 18 postsMetrics, logs and traces with OpenTelemetry and Grafana.
View all 22 postsInfrastructure as code with Terraform and friends.
View all 12 postsIdentity, network and workload security in the cloud.
View all 30 postsImages, containers and local container tooling.
View all 19 postsBuilding and running a home lab for hands-on cloud and Kubernetes practice.
View all 15 posts- My Homelab in 3D: Inside the DGX Spark and the Homelab PC
- A Virtual KVM for Two Servers with Input Leap (TLS-Encrypted, X11 + Wayland)
- My DGX Spark was holding 54 GB for models nobody was calling: scale-to-zero vLLM with KEDA
- Backstage SSO broke for 3 days because the portal started 21 seconds before Keycloak
- k3d's load balancer ran out of connections, and DNS took the whole cluster down
- Getting a Targus DisplayLink Dock Working on the NVIDIA DGX Spark (Ubuntu 24.04, X11)
Question banks for DevOps, cloud, security and architect roles.
View all 19 postsLatest posts
A Virtual KVM for Two Servers with Input Leap (TLS-Encrypted, X11 + Wayland)
How I turned a DGX Spark and a homelab desktop into a single keyboard/mouse setup with Input Leap — server config, mutual-TLS encryption, systemd auto-start, and the X11-vs-Wayland gotchas.
Predicting memory leaks from heap dumps
Engineers use heap dumps and AIOps to detect memory leaks before application failure. By analyzing retention paths from GC roots, they prevent system crashes and maintain stability.
Grafana 'No data' on every panel: a node_exporter stuck at its concurrency limit
Every Grafana panel said No data. The cause was an empty dashboard variable and a node_exporter returning 503 after hitting --web.max-requests.
My DGX Spark was holding 54 GB for models nobody was calling: scale-to-zero vLLM with KEDA
How KEDA and its HTTP add-on scale idle vLLM model servers to zero on a DGX Spark, hold requests through a cold start, and give the unified memory back.
Grafana stuck in Init:0/1: the initDashboards flag that never lets go
Why the Grafana Helm chart's initDashboards option can leave the pod in Init:0/1 forever, and the native sidecar fix that keeps dashboards ready at startup.
Backstage SSO broke for 3 days because the portal started 21 seconds before Keycloak
Why Backstage's OIDC sign-in returns a 500 forever after one failed Keycloak discovery at startup, how I traced it, and the init container that prevents it.
Job Hunter Agent: Interview Questions
Interview answers from architect, developer, DevOps and SRE angles, drawn from building an AI job-application agent with a manual approval gate.
Job Hunter Agent: Continuous Improvement
How Job Hunter Agent keeps getting better: what the data says, what using it taught me in week one, and the roadmap and exit plan that follow.
Job Hunter Agent: Governance and Compliance
Light governance for a personal AI agent: who decides, what the platform enforces automatically, licences, LinkedIn's terms, and the evidence that the approval gate holds.