Shadow databases are normal
Every growing team accumulates instances: a reporting replica from a migration, a SQLite file behind an internal tool, a MongoDB someone spun up for a prototype that quietly became load-bearing.
Discovery as a continuous process
Running discovery once at onboarding produces a snapshot that is wrong within a month. The agent re-scans its host on a schedule and surfaces new instances in the dashboard with their type, version and size.
From there, enabling backups and monitoring for a newly discovered database is a single toggle — which is the whole point of keeping the inventory current.
