ONTAP Discussions

Interpret Grafana Stats



I can see on our AFF 8080 EX high Backend Latency. What does this mean and more important how to improve this?

General, how can this happen? We use All Flash SSD 8080 EX MCC Storage with peaks 30 000 IOPs and WAFL consum 10 Cores and Latency is up-to 10ms.


We need do add more Disks to the System? Can we improve WAFL CPU usage with this Action?






Re: Interpret Grafana Stats

It's hard to say from this but is your system running hot? Are you having latency issues in general?


Here are a couple posts which might help, but I believe that refers to CPU d-blade or WAFL level CPU processing:




Here is a KB that explains it a bit: https://kb.netapp.com/app/answers/answer_view/a_id/1001217/loc/en_US. Basically it would correlate to WAFL_Exempt.

Re: Interpret Grafana Stats

yes we have. Up to 30ms for some Volumes (Workload ~ 26000 IOPs 80% read, 1.5GB/s )

For an AFF8080 this workload should not cause such latency impact, right?


What means qos_latency_from_backend und more interessting how to solve ?



Re: Interpret Grafana Stats

That's kind of hard to say without knowing your other things going on with that system and the type of IOPs.


Have you talked to your account team about this? Using qos statistics can be very helpful, and if you're really stuck go ahead and open a case.


A couple articles which might give a clue too: https://kb.netapp.com/app/answers/answer_view/a_id/1075556 https://www.netapp.com/us/media/tr-4211.pdf

Re: Interpret Grafana Stats

well, 80% read /random as well sequential/.  We use one volume for ora Data files with 150 dbf meanwhile.

Volume peak ~ 30000 IOPs. During peaks CPU UT 100%  (20 Cores Plattform)


I already opened Netapp Case.

Cloud Volumes ONTAP
Review Banner
All Community Forums