HealthScores are a good way to monitor faults and the general health of your ACI Fabric. But in certain cases, such as CRC errors, if the interface does not change state (up/down), no fault will be generated and consequently the HealthScore is not affected.
In this article we examine what steps are necessary to ensure that errors such as CRC, FCS affect will indeed affect the HealthScore of our ACI Fabric.
Goto Fabric > Fabric Policies > Monitoring Policies > default > Stats Collection Policies
Then select “Layer 1Physical Interface Configuration” and then select “Ingress Drop Packets” (as shown below).
Click on the “Config Thresholds” + symbol and at the new pop-up wizard, edit the thresholds for “Ingress Errors Drop Packets rate”
Modify the values as needed for Rising
This will count the packets drop (CRC) in a 5 min interval and raise a Fault when the threshold is hit and will affect the Health Score accordingly.
In the figure above, if there are 10 CRC errors in a 5 minute interval then a fault of Warning will be raised, if the count increases over 500 then a Minor fault will be raised and so on. You can change the counts as per your environment.