Hello, I need to create a safe SQLite online backup from a running Railway service with a persistent Volume.

Setup:

• Node.js service
• Persistent Volume mounted at /data
• Live SQLite database: /data/travelcalc.db
• SQLite WAL mode is active
• travelcalc.db-wal and travelcalc.db-shm currently exist

I need to create a transactionally consistent backup of this live SQLite database without restarting, redeploying, detaching, or modifying the Volume.

The Railway CLI SSH connection asks me to trust the host key for ssh.railway.com, but I cannot find an official Railway-published host-key fingerprint that I can independently verify.

For security reasons, I do not want to bypass SSH host verification or use StrictHostKeyChecking=no.

I can already access the Volume using:

railway volume files list /

and I confirmed that travelcalc.db, travelcalc.db-wal, and travelcalc.db-shm exist.

However, I do not want to download those files separately because I cannot confirm that Railway guarantees a consistent point-in-time snapshot across those files while SQLite is actively running.

Could Railway please clarify:

1. Is there an official/verifiable method to verify the SSH host key for ssh.railway.com?

2. Is there an officially supported way to execute a one-time command inside the currently running deployment without SSH?

3. What is Railway's recommended method for creating and downloading a transactionally consistent backup of a live SQLite database running in WAL mode on a Railway Volume?

4. Do Railway Volume Backups guarantee a SQLite/WAL-consistent point-in-time snapshot while the database is actively being written to?

My goal is simply to create a safe off-platform backup without changing the production database.

Thank you.