[Netdisco] SNMP Connect failures - source of IP?

[<on...@LE...>]

In the SNMP Connect Failures page are a number of IP addresses on a subnet I do not recognize.  They are current and showing connect failures, and indeed I can't ping them either.

We have hundreds of 510 devices; they are coming from somewhere amongst them, in a LLDP or CDP neighbor probably, maybe elsewhere.

Is there a way to tell where they originate?   Short of trying to run each device individually in debug?

Is a record kept of what causes a discovery job to be queued for an IP, that you can look retroactively?  Somewhere in the database is fine (I looked briefly but had no clue where to start).

Linwood