Because sysUpTimeSecs is a 32-bit counter of 1/100 seconds, it will overflow at 497.1 days and reset to zero, tripping thresholds that warn against a low uptime and potentially creating notifications based on those thresholds.
The patch evaluates snmpEngineTime and uses it instead of sysUpTimeSecs to avoid alerts on switches with long uptime. It can also be applied to a Cisco 7206VXR (7200 series) router and presumably other Cisco (and other companies') devices. The "message" file displays the difference between them but perhaps should be edited to be less of a debug statement.
More at https://sourceforge.net/mailarchive/message.php?msg_id=27149286
Patch to use snmpEngineTime and see difference from sysUpTimeSecs