Load config file from URL

[DiniFarb]

As requested in #1207 I've added support for loading the config file either from a directory or a URL.

[breed808]

Thanks for submitting this! Would it be possible to include a flag allowing a user to skip TLS verification for the config file? I know some users of the exporter may have self-signed certificates in place, despite it not being best practice.

Something like this, with insecure_skip_verify being passed to the NewResolver function as a bool parameter:

tr := &http.Transport{
        TLSClientConfig: &tls.Config{InsecureSkipVerify: insecure_skip_verify},
    }
    client := &http.Client{Transport: tr}
    _, err := client.Get(file)

[DiniFarb]

Good point and it is good to have such a flag for testing purposes as well. I'll implement that.

The flag should then be passed as an argument like:

--config.file.skip-tls-verify=true right?

Or somehow passed in as a prefix or suffix in the url? Like:

--config.file="[insecure_skip_verify]https://example.net/config.yaml"

[breed808]

Yes, a boolean flag would be preferable here (assuming the flag library supports that type).

[DiniFarb]

Done, now its possible to skip TLS.
One more question though, I saw that the --config.file flag is not listed in the global flags section of the readme. Wouldn't it be good to add the config flag plus the new skip TLS flag well as?

[breed808]

Sure, feel free to add it in the readme if you like.

One other thing: you'll need to sign-off the commit(s) in order to pass the CI DCO check. I can't recall if all commits need to be signed or just the last, but I'd assume the worst and sign all of them.

Let me know if you need assistance with that.

[DiniFarb]

I'am not quite sure, should I use git rebase --signoff HEAD~6 to sign commits afterwarts ?

I have added the readme change directly online so the last commit is signed.

[breed808]

Yes, the git rebase --signoff command should do it if the DCO check continues to fail

[breed808]

Hmm, the force push seems to have added other commits, and there's been a couple merge commits since then.
Want some assistance rebasing the feature branch? I believe I can push to it.

[DiniFarb]

Yes I have always synct the pr branch as soon as it was behind the prometheus-community:master. Shouldn't I have done that? 😬 sorry for that, and yes u should be able to push to the pr branch. Otherwise just write the rebase command I should use.

[breed808]

Syncing it with the upstream master branch is generally advised, the issue here is that this PR is proposing to merge your master branch into the master branch of the windows_exporter project, so we'd end up merging merge commits. Merges all the way down 😄

I'll force-push to your master branch now in order to tidy the commits in this PR, once I've pushed have a review of the commits and let me know if you're happy with the result. If not I can revert the changes.

[breed808]

Commits look alright now, though the DCO check is complaining again ☹️

Author: andreasvogt89, Committer: Ben Reedy; Expected "andreasvogt89 30302212+andreasvogt89@users.noreply.github.com", but got "Vogt Andreas, INI-MBM-BNC Andreas.Vogt3@swisscom.com"

[DiniFarb]

yes my bad, I've used the wrong account last week to sign the commits 😐 I've done it again, but with the fork sync I've fu* up the commit timeline again. Could you help me with the rebase ?

[breed808]

Got it, the DCO check is happy now. Have a review of the commits and let me know if you're happy with them, then I'll get this merged.

[DiniFarb]

I am very sorry for the confusion I've created here and many thx for fixing all 😊 The commits are looking very good and clean to me now 👍🏼

[breed808]

It's fine, nothing to worry about. git is not the easiest tool to use, especially when trying to cleanly merge changes.