Print disk usage information in ci logs

[asedeno]

Risk Level: low
Testing: ci-workflow change

[phlax]

@asedeno we have only just removed this in favour of printing to a report json artifact for each job - i think we dont want to add it back

any reason you cant use the json report?

[asedeno]

Why not both? An artifact is nice, but seeing an entry in the logs that say it's capturing free disk information and then not finding said information in the logs right there was very confusing.

[phlax]

Why not both?

its rare that you need this information and distracting when you are trying to figure out what broke

same argument could be made for the du debugging (as well as others)

[asedeno]

And yet, I ran into a full disk in #36699 and was very confused by this information being collected and not presented to me. Downloading a zip file, unpacking a json artifact, and pretty printing it to see something that could have just been in the logs felt like needless busywork for something that could have just been printed.

[phlax]

And yet, I ran into a full disk in #36699 and was very confused by this information being collected and not presented to me. Downloading a zip file, unpacking a json artifact, and pretty printing it to see something that could have just been in the logs felt like needless busywork for something that could have just been printed.

perhaps this can be resolved with some dev documentation - as explained offline this really doesnt give you any of the information you require to resolve - just a confirmation of what you already know

the download and unzip step is perhaps a little convoluted but tbh is pretty trivial with the help of jq - ive been doing it frequently while debugging the migration

[phlax]

... and the ci-report does give you the info you need

[asedeno]

I still think that this is useful information to print, but based on offline conversation, I'm closing this for now.