Network Observability Workshops

Self-contained, laptop-friendly workshops for Modern Network Observability — the book and the talks. Each workshop spins up a complete observability stack (Prometheus, Loki, Grafana, Alertmanager, Telegraf, Infrahub, Prefect) plus a synthetic telemetry generator (sonda) that stands in for a small network. No real or containerized network devices required.

Why this repo exists

The companion repo network-observability-lab grew out of the book's chapter exercises. It is excellent if you want depth — every chapter, every variant, every collector — but the surface area is large and the setup assumes you can run cEOS / SR Linux containers locally.

This repo is the opposite trade-off: one Taskfile, one docker compose up, one observability stack per workshop, and sonda doing the network-shaped telemetry instead of real devices.

Repo layout

Taskfile.yml             # umbrella tasks; namespaces into each workshop
pyproject.toml           # uv workspace root (members = packages/*, workshops/*)
.python-version          # Python version pin (uv reads this)
scripts/                 # shared shell helpers (workshop-agnostic)
  preflight.sh           # check docker, RAM, disk, network reachability
  sonda-trigger.sh       # generic sonda/Loki triggers (e.g. flap-interface)
packages/
  nobs/                  # shared Python CLI used by every workshop
                         #   nobs status / alerts / schema load / maintenance
notes/                   # local-only scratchpad — gitignored, see notes/README.md
workshops/
  autocon5/              # AutoCon5 — Modern Network Observability workshop
    README.md            # attendee-facing instructions
    Taskfile.yml         # workshop-scoped tasks
    pyproject.toml       # workshop's Python deps (uv workspace member)
    docker-compose.yml
    .env.example
    lab_vars.yml         # source-of-truth data fed into Infrahub
    sonda/               # synthetic telemetry scenarios + helper scripts
    prometheus/          # config + alert/recording rules
    loki/                # config + alert/recording rules
    alertmanager/        # routing config
    grafana/             # provisioning + dashboards
    telegraf/            # scrape config for sonda-server
    logstash/            # GELF -> Loki ingest
    infrahub/            # schema YAML
    scripts/             # workshop-specific helpers (Infrahub loader, set-maintenance)
    webhook/             # FastAPI receiver for Alertmanager
    automation/          # Prefect flows (alert -> evidence -> decision -> action -> RCA)

The umbrella Taskfile.yml namespaces tasks per workshop:

task setup                                          # uv sync (Python deps)
task preflight                                      # generic env check
task autocon5:up                                    # bring up the AutoCon5 stack
task autocon5:down
task autocon5:flap-interface DEVICE=srl1 INTERFACE=ethernet-1/1

Two CLIs are installed when you uv sync:

• nobs — generic operator toolkit (status, alerts, schema load, maintenance). Reusable across workshops and ad-hoc work. Lives at packages/nobs/.
• autocon5 — workshop CLI. Re-exports every nobs subcommand and adds workshop-specific ones (load-infrahub, evidence, try-it). Lives at workshops/autocon5/src/autocon5_cli/.

Attendees only need autocon5. nobs is exposed for anyone driving an arbitrary stack.

Prerequisites

• A laptop with Docker (or Docker Desktop / Colima / Rancher Desktop) and Docker Compose v2 — docker compose version should report v2+.
• go-task — brew install go-task on macOS, or grab a binary from the releases page.
• uv for the Python tooling (curl -LsSf https://astral.sh/uv/install.sh | sh). uv handles Python installation too, so you don't need a system Python.
• Git to clone the repo and a few GB of free disk for container images.
• Approximately 8 GB of free RAM while the stack is running.
• Outbound HTTPS to github.com, ghcr.io, docker.io, quay.io for image pulls.

Quickstart

git clone https://github.com/network-observability/workshops.git
cd workshops

# One-time: sync the Python helpers into a workspace .venv/.
task setup

# Generic environment check (works for any workshop).
task preflight

# Bring up the AutoCon5 stack.
task autocon5:up

# Seed the source-of-truth (Infrahub).
task autocon5:load-infrahub

# When you're done.
task autocon5:down

Each workshop's own README walks through the agenda and the hands-on parts.

Available workshops

Workshop	When	Outline
autocon5	AutoCon5, Mon 4 May 2026	Telemetry & queries → dashboards → alerts, automation, AI-assisted ops

License

MIT.