The Sequencer toolset is still being actively developed, so you may want to do some feasibility testing to see if it’ll work for you before you dive in. That being said, you can find the basic documentation on recording gameplay with Sequencer here.
You’re essentially looking at a two part process. First, you’d need to run Sequence Recorder to create a Level Sequence with all of the relevant actors. Then, you’d use Sequencer to render that level sequence out to a video file. Of course, I’m not sure this would catch any rendering artifact that the replay system wouldn’t, since we’re still recording actor properties and playing them back to render out the video.