A long running transaction will sometimes cause something that looks like this.
There are two transactions as follows
1. 10000 operations and no commit.
2. Another transaction that has a commit.
The last processed message is the commit from the second transaction, but the number of messages in the queue is 10000 since the first transaction hasn't been committed. But when you do a "qstatus" the backlog still is 0. Poster will look like this until a commit for the first transaction comes in.
The other possibility that would look like this is a commit followed by a long running DDL command. Poster will wait for the DDL to finish, the commit is the last command processed and any other messages will queue up while the DDL is processed. If poster is restarted it will probably return to the same situation.
This does not necessarily signal a problem.
The easiest way to find out what poster is working on now is to query Oracle.