fix: delete the 'runner_config_toml_rendereded' output variable

[JulianCBC]

Description

Due to calls to nonsensitive in case the data is sensitive or not calling the function in the other case, the code gets very complex. The output variable runner_config_toml_rendereded was introduced for easier debugging, but we still have the option to write the whole Runner configuration to a local file (debug.write_runner_config_to_file).

This "fixers" that issue by just deleting the output variable.

Closes #1018

Migrations required

Usually not. For debugging purposes, set the debug.write_runner_config_to_file to true. The configuration is written to a file on your local disk (directory debug/)

Verification

Only linted as this is only deleting an output variable.

[github-actions]

Hey @JulianCBC! 👋

Thank you for your contribution to the project. Please refer to the contribution rules for a quick overview of the process.

Make sure that this PR clearly explains:

• the problem being solved
• the best way a reviewer and you can test your changes

With submitting this PR you confirm that you hold the rights of the code added and agree that it will published under this LICENSE.

The following ChatOps commands are supported:

• /help: notifies a maintainer to help you out

Simply add a comment with the command in the first line. If you need to pass more information, separate it with a blank line from the command.

This message was generated automatically. You are welcome to improve it.

[kayman-mk]

The fix is definitely needed. But the code looks rather complicated: var.runner_worker_docker_machine_fleet.enable == true ? nonsensitive(local.template_runner_config) : local.template_runner_config. It should simply return the template. But we have to wrap it with nonsensitive sometimes. I understand, why this is needed.

Maybe we find another solution to export the Runner's configuration? This was the original intend of this output variable. But I guess they will all have this sensitivity problem.

[JulianCBC]

The fix is definitely needed. But the code looks rather complicated: var.runner_worker_docker_machine_fleet.enable == true ? nonsensitive(local.template_runner_config) : local.template_runner_config. It should simply return the template. But we have to wrap it with nonsensitive sometimes. I understand, why this is needed.

I completely agree: the goal of this pull request was to start the discussion and to that end I chose the "smallest" fix that wasn't completely insane or removing functionality. We definitely need a better solution than this.

Some options from my perspective:

1. Just delete this output - it's not something people should need day-to-day unless they're either paranoid or debugging the module
2. Only output something in this output if a flag is set - essentially turning it into a debug aid only
3. Move the nonsensitive() call up into the inputs to the template, making that always non-sensitive - this would mean that the private key is effectively leaked into the state without any attempt to flag it as being private
4. Pass the private key to the EC2 instance another way

[kayman-mk]

Voting for 2 and delete the output. The module writes the configuration to a local file. This should be sufficient for debugging.

Would be nice to get rid of sensitive/non sensitive. And I wouldn't consider this as a major change, as it is for debugging only.

[JulianCBC]

Thanks @kayman-mk, I've re-done this pull request to just delete the output variable.

[kayman-mk]

PR updated as we use it as commit message.

[kayman-mk]

Thanks for fixing this bug.

[JulianCBC]

Thanks @kayman-mk for your help in sorting this out and fixing up the PR message.