Labs

Labs

3 hands-on labs. Each one is runnable code you deploy on your own machine, a set of experiments that each do one thing, and real screenshots of what the dashboards showed when the lab was run. Pick a lab and start at part 1, or jump straight to its experiments.

Observability

Traces, metrics and logs. Instrument an app, follow a request, and learn to read the dashboards.

LabBuilt

eBPF Lab

Swap a k3d cluster's networking for Cilium and watch every flow in Hubble, get RED metrics and traces from an app with no instrumentation code, and catch (then kill) a shell's download in a container with Falco and Tetragon, all with eBPF on one DGX Spark.

Advanced about 2 hours k3d (lab-ebpf) on the homelab network, Cilium as the CNI 7 experiments
  1. Overview and Architecture
  2. Write the Code
  3. Set Up the Infrastructure
  4. Deploy
  5. Monitor
  6. Experiments
  7. Understand and Review
Experiments: See every flow without kube-proxy Drop a connection by identity RED metrics from an app with no instrumentation A trace across web, api and redis Catch a shell in a container Kill a download before it starts Delete the cluster and keep the data
Open lab
LabBuilt

OpenTelemetry End to End

Follow one click from the browser through nginx, Traefik, a FastAPI API, a backend with Redis and SQLite, and into traces, logs and metrics in Grafana, all on a local k3d cluster.

Intermediate about 90 minutes k3d (lab-otel) on the homelab network 6 experiments
  1. Overview and Architecture
  2. Write the Code
  3. Set Up the Infrastructure
  4. Deploy
  5. Monitor
  6. Experiments
  7. Understand and Review
Experiments: Reload the page Follow a request on a cache miss Compare a cache hit Send a burst of traffic Jump from a log line to its trace Take the backend down
Open lab

Containers & Kubernetes

Build images, run workloads on a local cluster, and watch scheduling, scaling and failures happen.

No labs here yet.

CI/CD & GitOps

Pipelines and reconcile loops you can run locally, break on purpose, and fix.

No labs here yet.

Cloud

Managed services on Azure, AWS or GCP, with the cost and teardown spelled out.

LabBuilt

DGX Spark to Azure Private Endpoint

Reach a private-only Microsoft Foundry account from an NVIDIA DGX Spark at home through a fourteen-dollar-a-month WireGuard VM, with split DNS that sends only the Azure AI names to Azure DNS, and watch every DNS answer and HTTP status flip in Grafana.

Intermediate about 90 minutes azure-cli + WireGuard (Azure VM) + docker-compose (local probe) 6 experiments
  1. Overview and Architecture
  2. Write the Code
  3. Set Up the Infrastructure
  4. Deploy
  5. Monitor
  6. Experiments
  7. Understand and Review
Experiments: Call the account over the internet Bring up the tunnel, change nothing else Send the Azure AI names through the tunnel Add the missing openai and services.ai zones Deallocate the VM to save money Start the VM and watch the tunnel come back
Open lab

Security

Scanning, policy and identity, shown through the signals they produce.

No labs here yet.

Data & AI

Databases, caches and LLM-backed services, observed end to end.

No labs here yet.