Storage Pool Utilization Reporting

[PFINCHNBCUNI]

I have several enhancements in with support already, but I believe this one would be valuable to almost everyone so I am adding this post.  I want the Storage Pool Utilization data by array over time to be ported to the DWH database so we can build reports using that data.  My reasoning for this is to generate reports with not only storage used, and IOP data, but also with percent average and max of the total performance that is possible in a storage pool.  I have a request in now with support, but know that the more of us that ask for this the sooner it will be worked on.  Please let me know if you need more information on why I want this or what I am asking for exactly.

Also, I write my own Dashboards and Reports, if any of you would like share or bounce ideas on these, please let me know.

Thanks,

Pete

[sunilyadav]

Hi Peter,

For what duration you are planning to generate that report?

And how are you merging or get utilization+performance data in same reports?

Please share your inputs and progress on case. I hope we see and learn something new here.

[PFINCHNBCUNI]

sunil-yadav

OCI is limited, but OCR should be able to report for 18 months.  I want as long as I can get.

[MARTIND82]

I have not yet started using DWH actively yet, but I definately agree seeing capacity and performance utilization over time would be extremely useful.  This would allow to show for growth or trend of utilization and performace over time. Kudos!

[ostiguy]

Hey guys,

Would using disk utilization numbers work as a proxy for storage pool performance utilization?

[PFINCHNBCUNI]

Yes, but we also need to report on the arrays internal cache to see if the box is overloaded but the disks are not. Both can cause latency issues.

[moechnig]

I don't think we collect stats about array cache now; am I missing an opportunity? 

This is going to be particularly important when we see more customers using Flash Pool aggregates on NetApp.  It would also let us track IOPS per TB of data and IOPS per TB of storage over time, which would enable some valuable analysis of workload characteristics compared to system capability. 

[PFINCHNBCUNI]

I think this is will become one of the most important metrics to track.  With all types of flash out now, overdriving the systems is becoming very easy to do, and harder to track down!!  OCI is a great tool but it needs to grow with what the arrays can do.  Thanks for jumping in.