Incident 1 (Editor/Primary): Visual interface instability (Express/Proxy) that did not interrupt data flow but generated header validation errors (X-Forwarded-For).

Incident 2 (8 days later): Total communication failure between n8n modules and Redis, resulting in ETIMEDOUT errors and the subsequent crash of all Webhooks.

Incident 3 (Another 8 days later): Database connection exhaustion or PostgreSQL unavailability.

Is anyone familiar with these types of failures or potential solutions? Before migrating to Railway, I used to host n8n and its dependencies on an isolated VM using Portainer, and I never experienced these recurring issues.