Provide API to inspect FTR state

[mshustov]

FTR config

It's hard to understand the final state of the FTR config due to its dynamic nature. Since it's not possible to determine the config shape with static analysis tools, we could provide API to show the FTR config value by request (it might be a CLI flag, an HTTP endpoint).
Input:

  return {
    testFiles: ...,
    services,
    servers: commonConfig.get('servers'),
    junit: ...,
    esTestCluster: {
      ...functionalConfig.get('esTestCluster'),
      serverArgs: ['xpack.security.enabled=false'],
    },
    kbnTestServer: {
      ...functionalConfig.get('kbnTestServer'),
      serverArgs: [
        ...functionalConfig.get('kbnTestServer.serverArgs'),
        '--server.xsrf.disableProtection=true',
      ],
    },
  };
}

Output:

{
    testFiles: ...,
    services: ...,
    servers: ...,
    junit: ...,
    esTestCluster: {
      license: ...,
      from: ....,
      serverArgs: ['xpack.security.enabled=false'],
      ...
    },
    kbnTestServer: {
      buildArgs:...,
      serverArgs: [
        '--no-base-path',
        '--server.xsrf.disableProtection=true',
        ...
      ],
      ...
    },
  };
}

FTR state

UiSettings

UiSettings are shared across the tests, it might be hard to enforce the proper state isolation, which can lead to barely debuggable problems. Like this #97222 (comment)
A developer can hit UiSettings HTTP endpoint to fetch the actual configuration when debugging a test, but it requires manual intervention to stop execution. Would it be possible to log the UiSetting state in debug logs?

Saved Object

Probably unnecessary. Large SO might add needless noise.

[elasticmachine]

Pinging @elastic/kibana-operations (Team:Operations)

[rashmivkulkarni]

cc @elastic/kibana-qa

[spalger]

I think something similar to the --print-config flag from ESLint would be nice here.

[wayneseymour]

Perhaps both are useful, a cli flag and an http endpoint.

[spalger]

I was confused about the HTTP endpoint suggestion, what would the server be in this case? FTR config being consumed by code that might use an HTTP endpoint seems like it would only be relevant to the FTR code, which is running in the same process and has access to the config service, where these values can be read without HTTP. Can you expound a bit on how an HTTP endpoint would be used?

[wayneseymour]

Oh yeah, I forgot about that bit. Good point.

…

On Wed, Apr 28, 2021 at 4:02 PM Spencer ***@***.***> wrote: I was confused about the HTTP endpoint suggestion, what would the server be in this case? FTR config seems like it would only be relevant to the FTR code, which is running in the same process and has access to the config service, where these values can be read without HTTP. Can you expound a bit on how an HTTP endpoint would be used? — You are receiving this because you are on a team that was mentioned. Reply to this email directly, view it on GitHub <#98311 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABFCBQY6AC5WY7HVTK7INMLTLCAXPANCNFSM43TCX54A> .

-- *Tre' Seymour* Senior Software Engineer Kibana QA - JavaScript Engineer in Test https://www.elastic.co/