Hi again Venoym,

Could I get your advice on an appropriate fix?

The functionality behind this service broker implementation is to log messages occuring from remote sites. Basically the remote sites log information about processing particular jobs back to the controlling server (via a webservice). These messages get stored in the database along with each job. I've tried with this 1000 batching mechanism to reduce the number of message records stored for each job, essentially the messages get concatenated and stored on one record, rather than one message per record.

But because of this behaviour, the batching it is not having the desired effect.

I was thinking, If I changed the conversation store, to store the ID of each job and grabbed the conversation for the job, rather than for the SPID, could that be a fix for this situation?