When the kubelet volume stats metrics will be supported in csi-vxflexos?

Hi Jian,

We don't expose metrics yet. We are working on a prototype for Prometheus integration.

Can you please contact me in PM so we can check your use-case and maybe prioritize that feature in the roadmap

Rgds.

Hi,

We are planning to implement threshold alarm for PVC usage too high which will rely on these metrics, and we found most of persistent storage types supported these metrics already, but vxflexos doesn't support it yet, so just check the plan for that.

Besides, whether vxflexos have similar function in storage server side? I mean the management system of vxflexos? Is it possible to define similar threshold alarm per PV in server side?

Thanks a lot for your support.

BRs/Jian

Hi Jian,

Here are the metrics we wanted to implement and expose 1st :

• Total capacity
• Use capacity
• Available capacity (worst case I can recompute it with the number above)
• Total number of volumes
• Read/Write Bandwidth (KB/s)
• IOPS
• Read/Write Latency (ms)

I will add the inodes to the backlog

Regarding the same VxFlex-side, you can configure from the GUI or the CLI a threshold on the capacity to trigger an SNMP alert (cf.  https://support.emc.com/docu67392_ScaleIO-2.0-User-Guide.pdf) § Setting a capacity alert threshold page 163).

To collect metrics, you can use the REST API call against /api/types/Device/instances ; the user guide documents the API usage. 

For a more polished/user-friendly experience, you can use the Monitoring and Reporting tool named Dell SRM : https://www.delltechnologies.com/en-us/storage/storage-resource-manager.htm

It goes much beyond the VxFlexOS and simple chart & alerts, it might be overkill for you.

Contact me in PM if you need further details on the options or you want to see an SRM demo.

--Florian