Application level healthcheck

[bdero]

We should have an application-level heartbeat route which can be used for health checking by monitoring services.

The body of the response should be machine readable and contain:

1. The status of each connected component of the application (databases, external application APIs, etc).
2. The elapsed time for each component health check to complete.

Return status of the health check request:

• 200 if everything is fine.
• 500 if one or more components is in a known degraded state (e.g. connection timed out).

Not strictly required for, but related to #10.

[bdero]

Lore has a good example of this: https://github.com/mitodl/lore/blob/master/status/views.py

[bdero]

One potential reusable solution is django-health-check, but it doesn't have:

1. Tests,
2. a secret token,
3. or a JSON/XML view.

It'd probably be easy to fork this and make it better if we want.