google / cadvisor

Analyzes resource usage and performance characteristics of running containers.
Other
17.09k stars 2.32k forks source link

Add I/O metric for persistentvolumeclaim and persistentvolume #3074

Open olahouze opened 2 years ago

olahouze commented 2 years ago

What would you like to be added: We would need to visualize the I/O that each pod performs on the PVCs it uses

We would also need to have the I/O of each PV

Why is this needed: We can currently monitor the I/O on the nodes' disks with container_fs_writes_bytes_total / container_fs_reads_bytes_total

But we can't identify which POD is doing the most I/O on a PVC We need to identify which PODs are doing the most I/O on a PVC

Describe the solution you'd like

Have the following metrics:

kubelet_volume_stats_io {endpoint="https-metrics", instance="xxx", job="kubelet", metrics_path="/metrics", namespace="xxx", node="xxx", persistentvolumeclaim="XXX",persistentvolume="XXX", service="xxx"}

Additional context If you already have a metric that allows you to do this and I missed it, thank you for your information

I don't know if I created this issue with good tag @bobbypage or @Creatone

olahouze commented 2 years ago

Hi

Could you take care of the ticket ?

Best regards

olahouze commented 1 year ago

Hi

Could you take care of the ticket ?

Best regards

ringerc commented 1 month ago

https://github.com/kubernetes/kube-state-metrics/issues/1701 and https://github.com/google/cadvisor/issues/3588

This needs to be done in another tool, kube-state-metrics doesn't have the info available.