Job runner details

[apj0nes]

Resolves #125
This change amends the job runner_type label logic as not all self-hosted runners may contain the self-hosted runs-on label. For example users of Warpbuild with have a runs-on label of warp-* but I would consider this self-hosted (non-github hosted runners). In addition to this, the runs-on labels have been added to job metrics as a new label (comma-separated).

This change should provide more insight into which runners are running which job.

Note: Github hosted runners list, this can be defined via config if the list becomes outdated and also allows users who user larger runners to add their runner labels.

[tcarmet]

Thank you for your contribution @apj0nes! It's very much appreciated.

Added a couple of comments check them out and let's make this PR move forward 🚀

Cheers!

[tcarmet]

Not sure how the doc was rendered (through mkdocs or github), through github at least couldn't get it to redirect unless I go up a couple of folders.

Also added some note for my future self as I was wondering how we could handle GitHub hosted larger runners with this setting.

Suggested change

	If the runner tag is in the [github_hosted_runner_labels](./gh_actions_exporter/config.py) list the runner type will be `github-hosted` else the runner type will be `self-hosted`. Additional `github-hosted` runner tags can be added via config.
	- If the runner tag is in the [github_hosted_runner_labels](../../gh_actions_exporter/config.py)
	list, the runner type will be `github-hosted` else the runner type will be `self-hosted`.
	- Additional `github-hosted` runner tags can be added via the config file,
	for example GitHub hosted larger runners as their name is defined
	by the owner of the GitHub organization.

[tcarmet]

What kind of query or filter were you trying to do when deciding to add the runner_labels and have you tried using the relabel option?

I see no issue in adding it, just wanna make sure you tried all the options available to you.

For example in our org, our self-hosted runners are defined with multiple labels:

• The os (ubuntu-20.04)
• The flavor (large)
• And in some case the cloud provider (aws, azure, etc)

And we use the relabel feature to retrieve those value and create new label like:

• cloud=aws
• flavor=large

Which makes it easier for us to query around.

[apj0nes]

This does look promising. Would you have an example yaml config for achieving the example above?

[tcarmet]

I do yes, this is an example of what we've been playing with:

job_relabelling:
  - label: cloud
    default: github-hosted
    values:
      - azure
      - aws
    type: name # retrieve the info from the runner name (defaults to labels).
  - label: image
    default: unknown
    values:
      - ubuntu-latest
      - ubuntu-20.04
      - ubuntu-22.04
      - rocky-8
      - redhat-8
      - redhat-9
  - label: flavor
    default: medium # consider github-hosted flavors as medium
    values:
      - small
      - medium
      - large
      - xlarge

And sorry for the lack of documentation on those 😅 , I'll take note on my end and make an effort towards that in the coming days.