Include buffering metrics in adapter health

In the OmfEgress component in adapter health the metrics are limited to DeviceStatus, and IO rate. Another key metric for this component is the buffering state. To keep it simple it would be useful to have the number of events in the buffer. This would allow for various remote monitoring use cases:

1. During a OMF Endpoint outage it would be possible to check that adapters are buffering data as expected, and if not an intervention could be made before significant data loss.

2. After an OMF Endpoint outage it would be possible to keep track of the buffer being purged, and understand how long until the adapter is back to real-time. This can be important for end-users to understand, and can be critical for analytics so that periodic analytics can be stopped until data is recovered and then calculation backfills can be carried out.

3. During normal operation it can be assumed that the number of buffered events should be stable - an increasing buffer queue could indicate an underlying problem that should be investigated.