Is there a VM Network demand metric that indicates that VM wanted more bandwidth than was available?

[MattG]

CPU and Memory demand show that those resources are under provisioned.   Is there a metric that will indicate that the VM is constrained on Network due to contention with other VMs or max NIC bandwidth.

Thanks,

-Matt

[MichaelRyom]

vRops doesnt know your acutal possible speed on net/storage of the VM/host and even if it did, it doesnt know if it is possible to get that speed end-to-end, on network and storage there is ussualy some fan-in/fan-out/oversubscription on ports end-to-end.

So its a though cookie for vRops - There are some calculated metrics and as I have been told those metrics are calculated based on the maximum throughput seen. This means that if you have a very low throughput, the calculations will be done on those and you at some point uses more this will be seen as your VM/hosts using all avaliable bandwidth untill a new normale is been established.

If you want here are some of the metric avalible:

Note the first three VMs in the list are all nested esxi hosts and the different "Current Size" and "Total Capacity" eventhough they have the same nics.

[MattG]

The thing is,  it could know if it wanted to as the connection speed properties are in ESXi.

-MattG

[sxnxr]

The problem with the connection speed properties that ESXi reported may not be the speed of the NIC. In our environment we use HP virtual connect flex fabrics. This allows me to carve up the 10 GB physical NIC into 4 separate flex nics that are presented to ESXi as physical nics (ESXi sees 4 Nics). I could have 1 NIC at 3GB, 1 at 1GB and 2 at 4GB but ESXi sees all 4 as being 10GB because of the way HP presents the speed to ESXi. This is one of the reasons vrops does not give you demand on the NIC as the presented speed is not always the same as the actual speed.