Data Connection Agent - Automatic Fingerprint Refresh

If a data connection source is setup via an agent, and the source’s fingerprint changes for whatever reason, I assume the syncs will just stop working, as the agent will not trust the source anymore?

However, if the old fingerprint is removed from known_hosts on the data connection agent’s host, would that be enough for the agent to make the connection, or is it required to manually SSH into the source system and approve the connection manually so that the new fingerprint gets saved? Some users have reported that the latter was the case for them, so I wanted to check if there are any guarantees/processes for Data Connection agents that account for these situations?