Adds HELP and TYPE to prometheus metrics

[djperrefort]

Resolves #308 by adding HELP and TYPE comments to the Prometheus outputs. An example is included below.

There is no help text provided by the health check framework. I considered using the instance docstrings, but no luck there either. I added the fully qualified plugin name as an alternative. This way we at least know where to go digging when there is a problem. I don't love this solution, so I could be easily convinced to do something else.

# HELP CeleryHealthCheckCelery health_check.contrib.celery.appsCeleryHealthCheckCelery
# TYPE CeleryHealthCheckCelery gauge
CeleryHealthCheckCelery{critical_service="True",message="unavailable: TimeoutError: The task took too long to return a result"} 500.0

# HELP CeleryPingHealthCheck health_check.contrib.celery_ping.backendsCeleryPingHealthCheck
# TYPE CeleryPingHealthCheck gauge
CeleryPingHealthCheck{critical_service="True",message="unavailable: Celery workers unavailable"} 500.0

# HELP DatabaseBackend health_check.db.backendsDatabaseBackend
# TYPE DatabaseBackend gauge
DatabaseBackend{critical_service="True",message="working"} 200.0

# HELP DefaultFileStorageHealthCheck health_check.storage.backendsDefaultFileStorageHealthCheck
# TYPE DefaultFileStorageHealthCheck gauge
DefaultFileStorageHealthCheck{critical_service="True",message="working"} 200.0

# HELP MigrationsHealthCheck health_check.contrib.migrations.backendsMigrationsHealthCheck
# TYPE MigrationsHealthCheck gauge
MigrationsHealthCheck{critical_service="True",message="working"} 200.0

# HELP RedisHealthCheck health_check.contrib.redis.backendsRedisHealthCheck
# TYPE RedisHealthCheck gauge
RedisHealthCheck{critical_service="True",message="working"} 200.0

[buttermutter]

lgtm