Replication struggling to keep up with 15 min RPO

[hostman]

Hi,

We have a deployment struggling to keep up with 15 min RPO, replicating a bunch of VMs from a vsphere cluster, SAN storage, Veeam running inside a VM on the vsphere infrastructure, replicating to a remote site.

We believe there may be some performance issues with the underlying SAN storage but not 100% sure at this point.

I know this is limited information, but can anyone suggest some basic tips / tweaks on improving the performance of this setup or some troubleshooting advice on pinpointing the issue (be it a Veeam issue, such as mis-configuration or some other underlying storage issue)?

[Shestakov]

Hi James,
what is shown as the bottleneck of the replication jobs?
Thanks!

[hostman]

Source is shown as the bottleneck.

There are 57 VMs in total on this infrastructure. Any other optimisation tips to reduce load on the source storage? The best we can achieve is 45 min RPO.

[Shestakov]

Could you provide the full bottleneck statistics, i.e. % for Source, Proxy, Network and Target?
Source bottleneck means the slowest part is the source disk reader reading the data from the storage. What transport mode do you use?