metrics: expose seastar runtime metric

[VladLazar]

Cover letter

Expose the "scheduler_runtime_ms" metric on the "public_metrics" endpoint.

Backport Required

• not a bug fix
• papercut/not impactful enough to backport
• v22.2.x
• v22.1.x
• v21.11.x

UX changes

• none

Release notes

• none

[mmaslankaprv]

lgtm, one question do we have uptime metric in the new endpoint ?

[VladLazar]

lgtm, one question do we have uptime metric in the new endpoint?

Good point. We don't, but we probably should. I'll create a PR.

[BenPope]

I was hoping to have the new metrics follow best practice of using base SI units, reporting in ms is disappointing.

[VladLazar]

I was hoping to have the new metrics follow best practice of using base SI units, reporting in ms is disappointing.

I agree it's not ideal. Seastar doesn't currently expose that through the public interface. We could patch it though.

[BenPope]

I was hoping to have the new metrics follow best practice of using base SI units, reporting in ms is disappointing.

I agree it's not ideal. Seastar doesn't currently expose that through the public interface. We could patch it though.

I'd rather add code to be able to map the value and push this feature to the next minor release than publish it with ms. We should do the right thing.

[VladLazar]

I was hoping to have the new metrics follow best practice of using base SI units, reporting in ms is disappointing.

I agree it's not ideal. Seastar doesn't currently expose that through the public interface. We could patch it though.

I'd rather add code to be able to map the value and push this feature to the next minor release than publish it with ms. We should do the right thing.

I hacked something together today. I'll create pull requests on Monday.

[VladLazar]

I created a couple PRs for reporting the same metric, but in seconds. It requires exposing the data from seastar first. If we find this approach preferable, we should revert this merge.

@BenPope have a look and let me know what you think. I've maked them as drafts for now to prevent them from being merged:

• scheduling_group: expose usage statistics seastar#31
• Publish runtime metric in seconds #5893