Recommended setup for 100 VMs

[Danne84]

I'am running a system that a need to make backups for 100 VMs during the night, But i'm having problems getting so many jobs to complete in so short time.
I have one proxy on each of the ESXi hosts (4 cores, 16gig ram on each)

The latest problem i have been getting is "active snapshots limit reached for datastore" and the jobs can stay in that state for several hours,
What is the recommendation for a system of that size?
Also need to "boost" the performance, Testing with jumbo frames (MTU 9000) but that does not seem to do anything.

The backup server is virtual and is on the same system that all the other VMs is that i'm backing up. (Using VMXNET3 NICs on all VMs)
The destination is an NAS with 16 drives in RAID 6 and connected via 10Gbit Fiber,

I'm grateful for any sugestions

[DGrinev]

Hi Danne,

Can you provide more information about the infrastructure and the settings:
1. Which transport mode is used.
2. How many backup jobs you have?
3. What is the bottleneck stats?
4. How many concurrent tasks allowed on the proxies?

Thanks!

[Danne84]

We are using VM proxys on each host so that means Virtual Appliance transport mode.
We have around 110 jobs
98% says target is the bottleneck, the rest says source and network.

8 concurrent tasks is allowed on each proxy.

[DGrinev]

I'd recommend you tremendously reduce the number of backup jobs. You will find useful information in the existing topic "Best way to structure backup jobs".
Also, you should look at your target storage performance, since the bottleneck is target.
Please can you share the transfer rate in Hotadd mode. Thanks!

[foggy]

What is the backup method used by the jobs (forward or reverse incremental)?