Trident Metrics Reporting

[acsulli]

Surfacing metrics surrounding Trident operations and functionality would benefit monitoring and other reporting for organizations which desire to track utilization. Importantly, this request is not for monitoring storage metrics, e.g. LUN/export IOPS or latency, but rather Trident specific actions such as:

• Number of volumes created/destroyed
• Volumes in use on a per-backend basis
• Time to provision/destroy volumes
• Storage array management interface connectivity/latency
• Combined with The ontap drivers should have more intelligent volume placement #64, reporting macro-level storage information which affects provisioning decisions (e.g. storage pool capacity used metrics which influence the decision engine).
• etc.

Andrew

[plz]

+1, thanks for capturing the requirement @acsulli

[ikruglov]

Hi! I would also propose:

• Errors during creation/destroying volumes? (at least just a counter)

[roldancer]

I recommend to define a Prometheus exporter for monitoring.

[vnandha]

Could we also add aggregator level metrics like ( capacity available and max volume count available? )

[msteenhu]

+1, I will look into Netapp api's and Trident integration, at least to know how hard it is... :)

[msteenhu]

I am interested in Kubernetes integration, there is a more focussed issue...

[peimanja]

Any update on this one?

[k8r-io]

Creating an prometheus exporter to even just expose basic health status would be a big step forward. Right now we tend to discover issues with trident when PVs stop provisioning.

[adkerr]

Added by 5da0122