How-to guides
How-to pages are task-oriented: each one gets a specific job done, assuming you already know roughly why you're doing it. For background understanding, use Concepts; for guided learning, use Tutorials.
The three flavours here
| Flavour | When to reach for it | Index |
|---|---|---|
| Operations guides | Planned work: deployments, onboarding, configuration, maintenance | CLI operator guide is the hub |
| Runbooks | Something specific must be executed precisely — often under pressure: incidents, game-days, recovery | Runbooks index |
| Recipes | Small, common developer tasks — add a strategy, run a backtest, query data | Recipes index |
Most-used pages
- Local setup — stand up the dev stack.
- Kubernetes deploy — the production rollout path.
- Incident response and kill-switch incident response.
- Rotate secrets.
- Add a strategy and run a backtest from YAML.
- Promote a bot to paper — pairs with the approval queues concept.
- Dataops: Airbyte + Dagster.
Conventions
- Steps are imperative and copy-pasteable; prose is kept to warnings and decision points.
- Pages state their prerequisites up front and link the concept page that explains the why.
- Anything an operator might run at 3 a.m. lives under Runbooks with explicit verification steps.