Add `Write-AzFuncLog` cmdlet to the helper module to allow writing Azure Function user logs in an explicit way.

This is a feedback from the tech review within PowerShell team.

Currently PowerShell built-in Write-* cmdlets are redirected to the Azure Function log in the following mapping:

Write-* cmdlet (the target stream)	Log level
Write-Verbose     (verbose stream)	Trace
Write-Progress    (progress stream)	Trace
Write-Debug       (debug stream)	Debug
Write-Information (information stream)	Information
Write-Host        (information stream)	Information
Write-Warning     (warning stream)	Warning
Write-Error       (error stream)	Error
Write-Output      (output stream)	Information

The log level Critical doesn’t have a corresponding mapping, and for Verbose, Progress, Host, and Output, it’s not obvious for users to tell what log levels they are mapped to. So the feedback is to provide another cmdlet to allow writing Azure Function user logs in an explicit way.

The proposal is Write-AzFuncLog, with the following syntax:

Write-AzFuncLog [-Message] <string> [-Level] <string> [<CommonParameters>]

As for the implementation, probably it should just call Write-Information -Tag internally with the log level as the tag.

[joeyaiello]

Per our discussion, we should talk to the Functions team about how critical is used in another languages.

/cc @asavaritayal @fabiocav

[daxian-dbw]

I don’t think it’s really… critical… personally though. Hence the request for feedback.

I think the critical concept in logging is determined by the business logic of a Function implementation. Maybe it’s not an error in script running, but the state of something makes a critical log necessary.

For number 2, the way I’m thinking is that, when I’m using Write-Verbose -Verbose in my script, I don’t necessarily mean to have my verbose message written to the trace level log. Instead, most of time, I just want to see them displayed to me truly as verbose messages. That’s why I think that proposal also makes sense to me. But, for Write-Error, Write-Information, Write-Debug and Write-Warning, their names match the log level so well that it would be confusing if they don’t really write to the corresponding logs. Hence, like I said, I’m not super sure about it.

Do you bombard someone with logs that can’t really be filtered like any other Function App?

Haven’t heard a user complaint yet, but we don’t have much usage today anyways.