Maintain and recover

Four pages for the life of an instance after go-live. The first three are procedures you plan; the fourth is for the day something is wrong.

• Promote between environments moves a service as one versioned package through export, lint, plan, apply and diff, and explains what a mid-apply failure leaves behind.
• Back up and restore covers the SQLite backup endpoint, the offline restore procedure, and what PostgreSQL and MySQL rely on instead.
• Upgrade an instance is the order that works for a version change: back up, preflight, validate the config, migrate, roll the fleet.
• Troubleshooting is indexed by symptom: what you see, why it happens, what to do, and how to confirm the fix.

Last verified 14 September 2026