GCP Server AutoDiscover: only log when the script fails

[marcoandredinis]

When running Auto Discover for GCP VMs, teleport always logs the output in debug level.

Now, it only logs the stdout/stderr when the installation fails. It also adds some context on which VMs it failed and the log level is now Error.

Related: #45172

[Tener]

This is logged in the loop over different IP addresses, so we may end up in a situation where subsequent attempt to a different IP address succeeds, but we logged an error-level information anyway. I think we should only log an error once we are sure we have ran out of IP addresses to try. So perhaps some kind of aggregate error?

[marcoandredinis]

This is logged in the loop over different IP addresses, so we may end up in a situation where subsequent attempt to a different IP address succeeds, but we logged an error-level information anyway. I think we should only log an error once we are sure we have ran out of IP addresses to try. So perhaps some kind of aggregate error?

I've change things slightly.
Now, we'll log only once in case of an error.
When the connection is successful with one of the IPs but the script fails, it will log the script stdout/stderr
When the connection is not successful, it will log all the errors returned (just like it was before, but now it uses Error instead of Debug).

[public-teleport-github-review-bot]

@marcoandredinis See the table below for backport results.

Branch	Result
branch/v15	Failed
branch/v16	Failed
branch/v17	Failed