By throttle, do you mean enabling network throttling, or are you limiting concurrent tasks (Proxy and Repository), or with Repositories..setting Read and Write latency metrics?

I think it’s network bandwidth throttling.  I created a Global Network Traffic Rule on the server and set throttling to 50Mb/s.

Ok...yep, that would be Network. Is your Repository storage local to the Windows 10 computer (i.e. direct-attached disk), or is it some external storage...for ex. external NAS or some storage, connected to your Windows 10 desktop as SMB, NFS, or other?

If external storage, then that would make sense your jobs would be running significantly less speeds with netwk throttling set.

The repository is local to the server (i.e., disks connected to the Win10 computer with the Veeam Backup & Replication installation).

Since my throttling is set to 50Mb/s, what’s confusing me is why my read speed would degrade all the way down to 5Mb/s or lower?

Hmm...not sure. I’m perplexed at that one too. Let me think about that, unless someone else chimes in with their thoughts.

The other thing that’s confusing is that when there’s no throttling, the Task Manager on the client computer shows Veeam Agent using around 97Mb/s, which is only around 10% of the theoretical limit of the installed gigabit ethernet adapter.  Why would browsers and email apps become unresponsive?  I know Veeam Agent is causing this because as soon as I kill Veeam Agent in Task Manager, Browsers and email apps work fine again.

I’ve read on the internet other users having this problem, but I couldn’t find a resolution.

Ok wait... I'm a bit confused at what you're using. Are you running Veeam Agent for Windows on this Win10 desktop or Veeam Backup & Replication Community Edition? Those are 2 separate products and both are not needed if you have them both installed on the same computer. 

I’m running Veeam B&R Community Edition on one Win10 computer.  The Community Edition manages backups on the second Win10 computer.  It installed the Veeam Agent for Windows onto that other computer automatically.

I re-enabled throttling.  As expected, read speed dropped back down to 4Mb/s.  But on the client computer, I see on the Task Manager Veeam Agent using 97Mb/s.  What’s going on? 

Ok, that sounds better.

Let’s take a step back to see maybe why that behavior is occurring. You really don’t need throttling on because everything is local. That is more a feature for latent networks, or to give more bandwidth to production loads over backup loads. I know it helps with the browser and app latency. But, really throttling isn’t the solution.

Let’s start with your Win10 recources, specifically what Veeam requires at a minimum to run as noted in the Guide. Do you have enough resources on your Win10 desktop to handle backups, even if it’s just 1 job? You require a minimum of 1 core and 4GB RAM for VBR. Using the Console requires an additional 2GB RAM; since you also use the default Proxy (i.e. the default Proxy installed with VBR on the desktop itself), that’s an additional 2 cores + 1 core per 2 concurrent tasks (disks) and 2GB more RAM + 500MB per concurrent task; then last, the Repository requires at least another core + 4GB RAM and additional one GB RAM per concurrent task. That totals to about 6 cores and 12GB RAM minimum, and that doesn’t include what the OS requires to run satisfactorily.

If you don’t meet those resource requirements, plus have additional resources to run OS apps/tasks, then there is where your browser/app latency is more than likely coming from when running your backup job.

When you throttle your network, which you are using network because the 2nd desktop you’re backing up is a remote computer, your backup speed will decrease because Veeam pulls data from the Source/backed up desktop to the target Repository, which when you don’t configure an external one, runs on the main desktop/computer Veeam is installed on.

When you run your backup job with throtlling and get low speeds, if you look at your backup job statistics (double-click the job), there are 4 metrics Veeam determines latency on: Source (source storage containing the data you wanna back up), Proxy (Veeam Component which performs all the data transfer, compression, dedup, etc. tasks), Network, then Target (Repository; storage location backups are written to). The metric with the highest number is where you’d start looking at for poor backup performance. I assume the Network metric would be the highest with throttling on. And that makes sense.

Anyway, that info hopefully should help. But, for your app latency during backups, I’d first look at your desktop resources to see if you meet those. If not, somehow add some if you can.

I assume you also have AV on these as well. If you have the default rules in place without any exceptions for Veeam this can cause a performance hit as well due to reading/writing of the data.

I don’t think it’s a resource issue.  The client computer having the problems is an older Dell Precision Workstation with a quad-core Xeon and 24G of RAM.  CPU and memory utilizations were relatively low when the browser and email client were having problems.

Instead of letting the Veeam server manage the backup, I tried letting the troubled client manage its own backup.  I deleted that client computer from the Veeam server infrastructure.  I then uninstalled the Veaam agent on the client computer (the one installed by the server) and reinstalled it manually from a download.  I initiated a backup from the client with no throttling, writing the backup files across the network to a shared folder on the same Veeam server.  I got read speed of around 58Mb/s with network utilization around 580Mb/s.  The were no responsiveness problem with browsers and email client.  Basically, my problem went away.  I’m not sure why having the Veeam server manage the same Veeam agent software on this particular computer would cause a problem with browsers and email clients.  I added two other Windows 11 computers to the Veeam server infrastructure.  The server installed the Veeam agent on those clomputers and manages their backups without issues.

Although have one computer manage its own backups is not optimal.  It’s an acceptable workaround for now until I can dedicate more time to figure out what’s going on. 

Hmm...well, glad to hear it worked for you..somewhat...with the workaround you implemented. So, using VBR to take Agent backups for other computers work ok?..but not this one particular computer? And, the apps/browser latency issue is on the client during backup, not VBR, correct?

Hi @ptmy305 -

Anything further you found to resolve this, or is your workaround of letting the Client do Standalone instead of Managed an ok resolution? Curious what your found, if anything.

Thanks.

My workaround is now my permanent solution.  It’s been working without issues for slightly over a month now.  Its good enough for me. 

Thanks for the update @ptmy305 

Best regards.