How to Validate Network Performance

We are a vendor for our customer and we have a flat /24 SCADA-like subnet on their network with all of our equipment. On that network I remotely guided them through upgrading 4 HP 1910 switches to HP 1920s models. After that I got the dreaded "now the network is slow" complaint. They reverted back to the old switches.

I can't imagine how the new switches could really cause a problem. There are other upgrades in progress on that /24 that are likely the true source of the problem. But at the same time I want to be humble and know that there are things I don't understand.

I want to get some metrics with the old switches in place and compare them to the new switches. Then present to management to prove it isn't the network. Below is what I plan to record before and after. Is there anything else I should document?

• document arp and LLDP entries on the old switches and compare them to the new ones. This way I can be sure they didn't mess up the cabling. The cabling is a 1 to 1 transfer but I can't rule this out as a problem

• iperf3 results to and from a sample of servers/workstations in the /24, before and after

• robocopy results of large ISO file (Windows environment obviously) to and from a sample of servers/workstation in the /24, before and after

• mtr stats before and after

• possibly put PTRG or LibreNMS monitoring into place?

Thanks for any advice!