Outbound Discards - Response Script?

Hello,

Looking for some ideas on how this might be possible! I have some Cisco equipment that has output discards due to egress buffer allocations. 


Optimization can be done when we know which interface require larger buffer allocations but we don't want to apply it to interfaces that do not need increased buffer allocation.

So far I use a Top XX Errors and Discards Today widget with the filter:
MachineType IN ('Cisco CGS-2520-16S-8PC', 'Cisco CGS-2520-24TC Connected Grid Switch')

The reason for limiting it to a specific model is because between different vendor or model there is sometimes a need for a different response so I like to look at one model or group of models at a time that share the exact same response.

I have a policy and class-map configured for CGS-2520 using the compliancy policy rules but the policy must be applied to any egress interface where there are discards and not ones where they are not having discards. To apply the configuration, you have to enter config terminal, the interface to be applied on and then apply the policy and save the configuration via:

conf t
interface (interface number that has output drops)
 service-policy output default-queue-limit
end
write memory

My question: any ideas how this could be automated through a job or remediation script on Solarwinds? It seems to require logic that is not available in Solarwinds but looking to see if there are any ideas. I need Solarwinds to detect the output discards on the interface and if there are discards then it should contain this configuration, if it doesn't contain the configuration ideally it would allow for it to be applied via remediation script or job.

For any interested the policy and class-map which is already configured and enforced through compliance policy is:
class-map match-all EF-class
 match ip dscp ef
!
policy-map default-queue-limit
 class EF-class
  priority
 class class-default
  queue-limit 544 packets

  • I've got a constant number on "Packets Received Discarded" for various Windows 2008 R2 Exchange mailbox servers in DAG.   Under Resource Monitor/Network/TCP Connections/Packet Loss shows "0".

    Shouldn't the number of lost packets fluctuate instead of a constant number such as 128, or 45,553?

    I" ve read many articles on increasing Rx Ring#1 and Receive Buffer.   Before that happens, I like to know if this is a false positive and how could I go about proofing it.