Scalability and performance troubleshooting

[SteelContainer]

Hi there,
I was just wondering if Veeam have any internal tools / scripts that can be used to assist with backup performance troubleshooting?
- Ability to list task slot usage across hyper-v hosts/proxies and repositories, usage of the slots over time or sort my most tasks in queue
- Ability to update hyper-v proxy tasks / CSV snapshot limits in bulk, cant seem to do this in powershell and tweaking settings of ~150 hosts is tedious via GUI
- Way to update the logging so it specify what resource it is waiting on - i.e. name the specific host or rep.
- Any other resource constraint reporting that goes deeper than the normal source > proxy > network > target percentages
- Any recommendations for larger hyper-v environments.

We have one environment in particular which has seen a slow down in backup performance around the start of April but it's difficult to find the exact cause and the informational messages in Veeam console/logs often don't narrow the issue down. It might say waiting on backup resources, rep, or proxy but not which one or how often this causes an issue.
This particular environment is approx 10 hyper-v clusters, ~150 nodes/hosts, 10 windows backup reps that range from 100-150TB ReFS 64K, 10G network.

I assume Veeam One can do some of these things but I don't know how much help it will really be.
I also have a support case open #04133657 to get some help there.

Appreciate any tips or suggestions.
Thanks!

[HannesK]

Hello,

1) I'm not aware of that
2) post358076.html#p357210 should help
3) the scheduler decides on the priority. In general: the earlier a job is started, the earlier it get's a task slot
4) I'm pretty sure, that this requires manual log diving. Even there, I'm not sure that you will find anything useful.
5) you already implemented the most important one: on-host proxy. The next thing is to reduce proxy tasks if Veeam requires too many resources or raise them if you see that backups are not influencing production

When the software says "waiting for resources", then it's about tasks slots. So checking the tasks slots on the repository could also help. It seems you have pretty many repository servers for the amount of hosts - so there should be enough resources. The message is also normal, when you start all jobs "at the same time". As said, the jobs are queued then.

Best regards,
Hannes

[SteelContainer]

Ok thanks for that.
I'll work with with support to see if we can tweak the settings.