Ok, after lots of digging in the code and @JJ-Fullmer database we seem to have found out what happend. Turned out some of his PowerShell code messed things up by deleting the primary MAC address of some hosts and the fog-client did its share to kind of reset some of the hosts’ information and make them pending.
The PowerShell code is fixed already and I will work on adding some checks to hopefully not mess as much up with the fog-client as we did here.
Initially we thought the snapin hash problems were at fault but the evidence collected points to this being a coincidence.