One of our open tickets on MidWest-IX is a member reporting slow speeds on their exchange port. After having them send us some data and a few e-mails back and forth we began looking at their switch port on the fabric. Right away we noticed errors on the port. After a counter reset the errors were still incrementing
19 runts 0 giants 1210 CRC 0 no buffer 1329 input error 0 short frame 0 overrun 0 underrun 0 ignored
This led us to look at our LibreNMS data for this port. A quick look shows on October 31st the port started seeing input errors.
By dilling down we are able to see exactly when this started happening
We now have responded to the customer to see if anything changed that day. Maybe a new switch, new optic, or software upgrade. By having this data available in an NMS we were able to cut down on troubleshooting by a huge margin. We now know when the issue started and are closer to the root cause of this. Without this data, we would be spending more time trying to diagnose and track down issues.