IOPS Numbers and Network Speed

Hi all! Apologies if this is asked and answered already. I was unable to find after numerous searches. Kindly point me to references if you are aware of them. I am not afraid to read, just can’t quite find the resource I am looking for. Basically, I don’t know what I don’t know, here.

My questions are related, but are separate. I can look at the reporting for the network bandwidth and the speed of my input/output. But how do I know what I am supposed to be looking for?

For instance, I am running two 10G Intel NICS connected via fiber to a 10G Arista switch. MTU is set to 9000 (across the network, physical and virtual switch, including the VMkernel adapters) and using iSCSI through the Arista to two Dell R730xd, also with dual 10G NICs, but connected to the Arista switch via DAC cables. The virtual hosts are configured for MPIO (two separate networks, two VMkernel adapters, two separate portals and configured on the targets, SATP RoundRobin claim rule set to 8 according to documentation). If I look at the reporting for one of the interfaces, I see the following.

Key Min Mean Max
Rx : 45.62 Mb/s 78.16 Mb/s 166.67 Mb/s
Tx : 4.19 Mb/s 7 Mb/s 53.91 Mb/s
Overlap : 4.19 Mb/s 7 Mb/s 53.91 Mb/s

Both interfaces are about the same (due to MPIO), so I only posted the one. The question here is if this is good or not? I wouldn’t know if this is good, bad, terrible, or great. When using the system, it feels responsive, but it could be terrible compared to another system. I am looking for some way to have concrete evidence I can use to set a baseline. Hopefully I am getting my question across.

I know I am not hammering the storage with constant I/O and assume the virtualization traffic is a lot of small files. But what I don’t know is if this is good, bad, or great? What do I use for a reference point?

Similarly, regarding disk I/O, I am have eight NETAPP X425_HCBEP1T2A10 (10,000 RPM) disks in an eight disk stripe. I am not concerned with data loss as this is a home lab, not-for-production, set up (it’s for learning and is constantly being built up and torn down; I can withstand a little downtime to replace components, if needed). I was more concerned with having more spindles for read write than if I had configured an eight-disk data vdev ZFS2 pool.
How do I know if I am getting good I/O?
Since I am running a stripe, almost all the disks report the same metrics.

Key Min Mean Max
Read : 253.06 KiB 582.29 KiB 1.13 MiB
Write : 407.4 KiB 790.55 KiB 2.38 MiB

Again, please excuse my elementary questions. My signature should contain my hardware. Please ask if I left anything out and will be happy to reply back with the relevant information.

Thank you for taking the time to read and thank you in advance if you reply!

-AR