Quickstart

The very first time, in this order:

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

uv sync --all-packages          # install workspace deps into .venv/
source .venv/bin/activate       # put nobs on PATH for the rest of the session
nobs setup                      # bootstrap .env + preflight
nobs autocon5 up                # first run pulls images, ~5–10 min
nobs autocon5 status            # repeat until every row says 'ok'
nobs autocon5 load-infrahub
# ... when you're done ...
nobs autocon5 down              # stop everything but keep volumes
nobs autocon5 destroy           # full reset (drops volumes too)

Activating .venv/ is what lets you drop the uv run prefix — nobs lives in .venv/bin/ once uv sync finishes. If you'd rather skip the activation step, prefix every command with uv run (uv run nobs setup, uv run nobs autocon5 up, ...) and the rest of this guide reads the same.

From inside a workshop directory the workshop name is optional too: cd workshops/autocon5 && nobs up is the same as nobs autocon5 up.

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

What you should see

After nobs autocon5 up finishes, run nobs autocon5 status until every row reads ok. First boot pulls 3–5 GB of images, so the first run can take 5–10 minutes; subsequent restarts are seconds.

Service              State   Health
─────────────────────────────────────
sonda-server         up      ok
sonda-srl1           up      ok
telegraf-srl1        up      ok
telegraf-srl2        up      ok
prometheus           up      ok
loki                 up      ok
grafana              up      ok
alertmanager         up      ok
infrahub-server      up      ok
prefect-server       up      ok
…

If any row is yellow or red after a couple of minutes, check the troubleshooting page — common cases (Infrahub still seeding, OrbStack proxy, slow image pull) all have one-line fixes.

---

Once everything is ok, head to the workshop overview for the four-hour walkthrough.