Add healthcheck endpoint to backend and frontend.

[plor]

Add a healthcheck endpoint that reports on whether connections (database, etc) are up and working. This should return 200 on healthy and appropriate error code on not-healthy. Also a JSON response should break down what components were looked at.

As a bonus add an caution state that hints that the service is not unhealthy, but may be having issues (this could inform scaling to add a new node or something similar).