Return suggested upgrade version from check_runner_upgrade_status

The following discussion from !90695 (merged) should be addressed:

• @ayufan started a discussion: (+3 comments)

  @pedropombeiro

  I find this logic highly complex and hard to understand what and for what case we propose those resolutions. I tried to rewrite to make this logic more flat. This makes all existing specs to pass correctly.

  TBH It would help to return also a proposed version to validate what is being proposed and when.

  
        def check_runner_upgrade_status(runner_version)
          return :invalid unless runner_version
  
          releases = RunnerReleases.instance.releases
          orig_runner_version = runner_version
          runner_version = ::Gitlab::VersionInfo.parse(runner_version) unless runner_version.is_a?(::Gitlab::VersionInfo)
  
          raise ArgumentError, "'#{orig_runner_version}' is not a valid version" unless runner_version.valid?
  
          # Recommend patch update if there's a newer release in a same minor branch as runner
          releases.each do |available_release|
            if available_release.same_minor_version?(runner_version) && available_release > runner_version
              return :recommended
            end
          end
  
          # Recommend update if outside of backport window
          if outside_backport_window?(runner_version, releases)
            return :recommended
          end
  
          # Consider to update if there's a newer release within the currently deployed GitLab version
          releases.each do |available_release|
            if available_release.same_minor_version?(@gitlab_version) && available_release > runner_version
              return :available
            end
          end
  
          :not_available
        end