The conductor component uses event-driven architecture to facilitate software automation.
It also maintains state/load-balancing of the runners and pipeline queue it is responible for.
Can the conductor service be on the same machine as runner service(s)?
Sure. But you are competing for resources.
Events are defined by you.
This allows events to also be defined through other 1st party services like the issue-tracker and merge-manager but also through 3rd party services like webhooks from GitHub or Slack.
Some events, like code automation, are ready to be wired-up out of the box, like git hooks that can be enabled through setting up new git code repositories (see here).
Other scaffold-ready 1st party events can be read about in the component documentation.
Events can be daisy-chained by calling the queue endpoint from your pipeline.