Maybe I'm misunderstanding it but what's the difference to using Grafana / Prometheus / Alertmanager?

I see the tool itself exposes the Prometheus metrics endpoint so I guess Prometheus already exists in your stack.

While Grafana allows you to monitor metrics provided by Prometheus and Alertmanager allows you to configure alerts also based on these metrics, Gatus allows you to directly monitor your service(s) by sending requests to them at interval and evaluating the conditions that make your service(s) "healthy".

As mentioned in the README, Gatus can double as automated user acceptance tests to check if, for instance, an endpoint that you expect to return X has Y instead.

Let me know if that's not clear enough, I can provide more a in-depth example :)