I like it a lot - it makes a runbook less dependent on external systems.
I hope that a future version of that block will also support a table view and perhaps even allow access to the results as a variable? I have a runbook for replacing the busiest worker in a Kubernetes cluster, and while it would be nice to see which one is the busiest in a table or graph, having its ID in a variable would make this even more useful.
One odd thing about the current Prometheus graph is the auto-refresh - the whole graph gets redrawn very visibly, like in slow motion.