Containers telemetry

[baronfel]

Fixes dotnet/sdk-container-builds#539

Adds telemetry for base image inference, container publishing, and a variety of errors that can occur along the process.

• add success telemetry
• add error telemetry
• add telemetry to allowlist in SDK's MSBuild logger
• verify telemetry is ingested
• classify new data points
• document new data points in the CLI docs/Containers docs

[baronfel]

Ok, I'm pretty happy with this now. We now send telemetry for container publishes for three main scenarios:

• container inference - to help us diagnose our decisions during inference
• container publish success - to know more about what kinds of registries and stores the images are being pushed to and pulled from, to help us diagnose and prioritize fixes to those different stores
• container publish failures - to help bucket the kinds of failures (authentication, logic error, layer digest errors, etc) to help triage more effectively and correlate errors with the kind of storage (in case some storages are more failure prone than others and might need retries, etc.

I've verified the data is allowed through the CLI's allowlist mechanisms, but haven't yet been able to verify it showing up in our telemetry systems.

All telemetry adheres to the CLI's existing opt-out mechanisms and we do not long user-provided information - we only log registry, image name, and tag information when we can verify that it's using the mcr.microsoft.com base images. In all other cases that data is null.

I'd like to get this in this week, because the 8.0.4xx freeze occurs soon.

[nagilson]

It seems wise to collect this information so we can better understand customer needs. I have mostly questions but it looks good to me otherwise :)

[baronfel]

I've verified that the telemetry data is being sent correctly - after merge GDPR classification will need to occur to get this 'done' done.