How-to guides
How-to guides are recipes for a specific job. They assume you know the basics (if not, start with the tutorials) and they skip explanations: follow the links to the explanation pages for the why, and to the reference for every option.
Install and run
- Self-host in production: requirements, install options, the production checklist, upgrades and rollbacks
- Deploy on Railway
- Deploy on Fly.io
Connect your data
- Connect a log store: Loki, Axiom, CloudWatch, Google Cloud Logging
- Connect metrics: Prometheus-compatible, Fly.io, Google Cloud Monitoring
- Connect deploys and commits from GitHub
- Connect Vercel: logs and deploys, through Vercel’s MCP server
- Ship Fly.io logs to a log store
Describe your system
Alerts and reports
- Send alerts to stromboli: openstatus, Alertmanager, Grafana, anything else
- Set up Slack
Tune investigations
- Choose models: providers, per-phase models, effort
- Enable health checks for problems that never fire an alert
Let stromboli act
Integrate
- Use stromboli from an MCP client: Claude Code, Claude Desktop, Cursor
- Script the CLI: CI, JSON output, exit codes
When something is wrong
Work on stromboli
- Develop stromboli: run from source, test, regenerate code