improve journal metrics and retention

if you stop processing in graylog and the journal get filled and will drop the oldest message by time or if space limit is reached.

```
# Journal hold messages before they could be written to Elasticsearch.
# For a maximum of 12 hours or 5 GB whichever happens first.
# During normal operation the journal will be smaller.
#message_journal_max_age = 12h
message_journal_max_size = 1gb
```

But the statistics of the node, means the message count will not show the drop of the messages. Only if you proceed your processing the message count will suddenly drop to the real value.
### ideas
- show a warning if messages in the queue will be dropped or get dropped
- display the correct message count
- calculate the right amount of messages in the journal
### Environment
- Graylog Version: 2.0.X