How to collect memory/cache bandwidth usage using cAdvisor?

[nfsp3k]

Hi,

I want to scrape memory/cache bandwidth usage by containers using cAdvisor on Kubernetes.
How can I do this?

My worker node has the flags cqm_mbm_total, cqm_mbm_local, and rdt_a set.

And resctrl is also mounted as follows.

However, when I try to scrape stats by using the REST API url 'xxx.xxx.xxx.xxx:8080/api/v1.3/subcontainers',
the result does not contain any metrics about memory/cache bandwidth usage.

I'm new to cAdvisor and I believe your response will be helpful to me.
My cAdvisor version is v0.30.2.

Thank you

[Creatone]

The Resctrl metrics were introduced in v0.37.0. Unfortunately only for "/" container (the entire machine).
#2793 PR adds metrics per container. I hope it'll be merged soon. Can you check if it's working for you?

[nfsp3k]

The Resctrl metrics were introduced in v0.37.0. Unfortunately only for "/" container (the entire machine).
#2793 PR adds metrics per container. I hope it'll be merged soon. Can you check if it's working for you?

Thanks for your reply, @Creatone. I'm going to check the PR.