Solarwinds is periodically reporting there are suddenly thousands or hundreds of thousands of errors on an interface. This is happening on multiple devices. It's as if Solarwinds is forgetting or losing the interface error count then when the next poll happens it reports the number of errors since the device was last rebooted (which for some devices is years).
The attachment is of a particular device that had false alerts several days in a row. Every one of those is the same number of errors, the errors aren't incrementing. Now, I know Solarwinds pulls the mib values, and I have done a snmp walk and checked that the interface error counts in the mib match what Solarwinds is reporting. So, that's accurate at least.
Why is this happening? And more so, what can I do to stop, correct, or avoid this?