Loss of interface Indices

We had a event, where a device was not reachable for one hour via snmp - from libreNMS point of view.

After aprox. one hour libreNMS was able to discover the device again - with new Interface indices, consequently we lost the RRD Data prior to this event.

Logs from start and end of this event:

2024-06-28 03:04:34 - Device status changed to Up from snmp check by 62d999bb613ea.
2024-06-28 02:09:02 - Device status changed to Down from snmp check by 62d99fd0603db.

The Poller 62d99fd0603db was never overloaded.

Any ideas how this can happen?
BR Adrian