We need to make a consistent offline copy of a SQLite database and its

associated WAL/journal files from a Railway Volume.

We checked Railway CLI v5.44.1. Its volume files download command requires

an active deployment and transfers files through the service instance.

Therefore, it does not appear to provide access after all application

instances have stopped.

Before we approve any downtime, could you confirm:

1. Is there an officially supported way to download selected files while

   all application instances are stopped, without deploying a helper or

   moving/replacing the original Volume?

   If not, please explicitly confirm this and identify the smallest

   supported alternative that preserves the original Volume.

2. For that supported method, does Railway start a temporary worker or

   mount the Volume? What are the source write/metadata effects,

   permissions, consistency guarantees, and automatic retry behavior?

3. What supported procedure confirms platform-managed instances have

   terminated, prevents automatic restarts/redeployments during the copy,

   and restores the exact previously deployed image afterward?

   Please distinguish restart policy from GitHub autodeploy behavior.

We understand that application-level SQLite writer shutdown must also

be verified separately on our side.

This is a request for guidance only. Please do not stop/restart services,

create a backup, modify files/settings, or detach/restore/move the Volume.