Print the porter version from the agent

[carolynvs]

What does this change

The porter agent used to print the porter version before executing the requested command. When I rewrote the bash script as a go program, that regressed and it stopped printing it out.

This adds that back.

What issue does it fix

I noticed when testing the operator that this broke. It's mostly useful for troubleshooting since the image in the porter agent pod is by digest, not a tag that you could figure out the version from.

Notes for the reviewer

N/A

Checklist

• Did you write tests?
• Did you write documentation?
• Did you change porter.yaml or a storage document record? Update the corresponding schema file.
• If this is your first pull request, please add your name to the bottom of our Contributors list. Thank you for making Porter better! 🙇‍♀️

Reviewer Checklist

• Comment with /azp run test-porter-release if a magefile or build script was modified
• Comment with /azp run porter-integration if it's a non-trivial PR

[VinozzZ]

Just curious, why do we want to send all non-bundle output to stderr?

[carolynvs]

In general, I've learned that it helps to decide what a user should expect to see in stdout vs stderr up-front. For example, outputting a mix of debug into and the actual command output (like a json document) to stdout makes it impossible to parse. We are still correcting that in porter in some places (as part of our instrumentation efforts).

Thinking about what a user of the operator would expect to see in stdout for the porter agent container, I am assuming that they want the output of the requested porter command. So even though the agent is doing a few extra things, like here we are printing porter version for debugging, it makes it easier for someone to separate out debug info vs the actual command output if we send porter version to stderr.

Let me know if you see a problem with sending the output of version to stderr for this container!