Skip to content

Administrator guide

Advanced operations

Run OpenCircle after first deployment without losing state, custody, or the thread.

Day-two work has one rule: operate the instance that already owns the data. Never create a fresh deployment just because an upgrade, claim, or browser handoff was interrupted.

Protect connections and secretsKeep durable custody outside process environments, logs, and source control.Upgrade safelyBack up, apply an immutable image, accept the result, or roll back.Back up and restoreVerify a recovery bundle before the day you actually need it.Diagnose a ServerSeparate application health, ownership, provider capability, and host authority.Railway automationDrive Railway through a structured, noninteractive command contract.Railway upgrade and recoveryProtect recovery custody and run the integrated Railway maintenance transaction.

Railway backup and recovery are phases of one host-upgrade transaction, not independent commands — the transaction can create and verify a replacement project when an in-place upgrade fails, but there is no general fresh-project restore entry point.