We run it, so you do not have to.
A forge you own should not become a server you babysit. These are the running parts we take off your plate, and what we commit to on each one.
Updates
We track Forgejo releases and apply them for you, security fixes first. You stay current without scheduling a maintenance window or reading a changelog to find out what broke.
Backups
Nightly and automatic, encrypted before they leave your instance and kept offsite with a rolling 90-day retention. Each forge backs up to its own isolated store, and restore is operator-run and follows a documented runbook we are still validating end to end, so we make no recovery-time promise yet. Snapshots are included on Team and Pro.
TLS
Certificates are provisioned and renewed automatically. Your forge is HTTPS from the first request, with no cron job to forget and no expiry to wake up to.
Runners
CI runners are provisioned, registered, and kept patched for you. Every plan runs on a pool we operate, and Team and Pro can attach runners dedicated to your forge. You write the workflows; we keep the machines that run them healthy.
Maintenance philosophy
Security fixes first
We track upstream Forgejo releases and prioritize security fixes over feature updates. Emergency work is handled as operational work, not as a customer-scheduled project.
Routine changes stay boring
Routine upgrades are batched, reviewed, and applied by Fjord. Where a release may affect customer behavior, we expect to communicate before relying on it in production.
Restore is part of operations
A backup is not treated as useful until the restore path is understood. Restore remains operator-run while we validate the runbook end to end.
Incidents get written down
Customer-impacting incidents are published with timeline, impact, cause, remediation, and prevention notes on the incident history page.
Ownership without the ops
The point of Fjord is to give you a forge you own without handing you a pager. We run the parts that have nothing to do with your code and everything to do with keeping it available: patching, renewing, backing up, restoring. And none of it is a trap, because you can export your data any time in Forgejo's own format, and if a plan lapses your backups are retained for up to 90 days (per our retention policy), so leaving never means losing your history overnight. If we ever stopped running it well, leaving would be the easy part.