My n8n instance on Railway has been inaccessible since this morning. Can't log in despite multiple reset attempts. This is critical as I have a client solution that needs to be deployed ASAP.

Setup:

• n8n self-hosted on Railway (PostgreSQL + Redis + Queue workers)
• 7 active production workflows (RAG, webhooks, client automations)
• Cannot truncate: I cannot afford to lose the workflows

What I've already tried:

• Deleted and recreated user in database with bcrypt hash
• Verified environment variables (DB_ , N8N_)
• Tested in private browsing / different browser / 4G network
• Cleared browser cache completely
• Redeployed n8n service (x5)
• Checked logs: no errors, no trace of login attempts
• Created a 2nd user with different credentials: same result
• Checked for intrusion: no suspicious users or IPs in database

Symptoms:

• Login screen displays correctly
• "Wrong username or password" on every attempt
• No logs appear when I try to log in (as if the request isn't reaching the DB)
• "Forgot password" doesn't work (no SMTP configured)

Hypotheses:

• Issue with N8N_ENCRYPTION_KEY that may have changed
• bcrypt hash not compatible with n8n version
• Frontend session issue

Question: How can I regain access WITHOUT losing the workflows in the database? Is there an API endpoint or CLI command to force a clean password reset?

Context: I have an urgent client delivery, so any quick solution would be greatly appreciated.