At the moment automations cannot be monitored for failures via Monitoring Views.
However, by default automations do send emails to the relevant owners when they fail. The set of users who get these emails are intentionally limited to prevent leaking data about the object types being automated
Not sure if this has been fixed, but I set up some logic to this.
It all comes down to your use case, but the basic flow is this:
Something happens that should trigger an automation (e.g. an Object is created/updated)
An automation is supposed to run because of this, with an estimated run time.
A ‘meta health check’ is run based on this.
This is obviously very vague, because the implementation sort of dictates what should trigger this.
It is also very possible, that you can solve it by setting up a second automation where the monitoring type is set to “Automation Dependency”. This will run it after the first automation, and you can then create some logic to check whether or not the first automation failed.
You can take this a step further if you make use of the Foundry API, to get things like build status, etc.
Not a first-class solution, but I’ve found this approach to be pretty solid.