My boss insists that our backup repo is somehow to blame but every backup run shows "Source" for the bottleneck. Rather than convince him of this I'd rather just work to improve throughput in general. Here's output from a recent incremental run of our largest job - do these speeds seem reasonable and if not where should I start looking? We've got everything on LACP-bonded 10-gig NICs, and our ESX vmstore is a Tintri.
No, frankly speaking these speeds are quite slow for a largest job with many VMs. I saw about same speeds 10 years ago on my laptop lab with virtual ESXi hosts running in VMware Workstation
You should do a webex with our support and see if they can spot any misconfigurations on your end.