Skip to content

Latest commit

 

History

History
30 lines (18 loc) · 2.8 KB

metrics.md

File metadata and controls

30 lines (18 loc) · 2.8 KB
Raw

Metrics are useful for debugging and monitoring the application in real-time. dynamic-nfs-provisioner uses Prometheus to report the metrics. dynamic-nfs-provisioner doesn't persist its metrics. If the dynamic-nfs-provisioner pod restarts, metrics will be reset.

By default, Dynamic-nfs-provisioner exposes the available metrics at path <POD_IP>:9500/metrics. You can change the metrics path by setting metrics-path and listen-address arguments in the pod template.

You can collect the metrics using cURL request to endpoint <POD_IP>:9500/metrics.

Follow the Prometheus getting started doc to spin up a Prometheus server to collect the dynamic-nfs-provisioner metrics.

Dynamic-nfs-provisioner metrics name has an nfs_volume_provisioner prefix as its namespace and subsystem prefix persistentvolume.

Metrics under persistentvolume subsystem describe the status of provisioner. To detect the failure in volume provisioning, These metrics should be monitored closely.

The following table lists all the metrics available under persistentvolume subsystem:

Name Description
create_total Total number of persistent volumes created
create_failed_total Total number of persistent volume creation failed attempts
delete_total Total number of persistent volumes deleted
delete_failed_total Total number of persistent volume delete failed attempts
  • create_total records the total number of successfully provisioned volume requests. This counter should increase over the time if dynamic-nfs-provisioner remains in healthy condition.
  • create_failed_total record the total number of failed provisioning request. This counter indicates temporary failure in provisioning because of invalid requests. Rising create_failed_total indicates the dynamic-provisioner is unable to serve the provisioning request because of the issue in provisioner or cluster health.
  • delete_total records the total number of successfully de-provisioned volume requests. This counter should increase over the time if dynamic-nfs-provisioner remains in healthy condition.
  • delete_failed_total records the total number of failed de-provisioning requests. This counter indicates temporary failure in de-provisioning because of invalid requests. Rising delete_failed_total indicates the dynamic-provisioner is unable to serve the de-provisioning request because of the issue in provisioner or cluster health.

To get the nfs server statistics, you can use the node_exporter. node_exporter exposes the nfs client metrics through collector nfs and nfs server metrics through collector nfsd. A detailed guide on how to install node_exporter can be found here.