<p>Excellent, thank you David.</p><p> </p>

<p>These look very helpful, if only i knew what they were telling me.</p><p> It looks obvoius but the devil is in the details, can anyone let me know how they work.</p><p>What does it do with jobs killed by daemons / filter rules?</p><p>What triggers the 'complete' counter reset ?</p>

[quote user="dilberts_left_nut"]<p>These look very helpful, if only i knew what they were telling me.</p>[/quote] It tells you how many jobs are in the queue. Jobs are divided into two classes - outgoing mail, and all others (includes internal deliveries, and mail from the outside being delivered to local users). A job in either class has three possible states: 1 - Ready:  The job is ready to be processed and will be attempted on the next polling cycle. 2 - Pending:  The job is in a wait state, waiting until a particular amount of time has elapsed. A job that had a temporary failure of some kind and has been rescheduled until a later time will appear as "pending". "Pending" is a completely normal state for outgoing mail, but is considerably less common for other classes (and may indicate a problem with the server or your user directory). 3: Complete:  The job has been finished and is waiting to be removed from the system. Note that a "complete" job may have succeeded, or it may have failed, but in either case, it means that every deliverable address in the message has had a formal result of one kind or another. For a variety of reasons, Mercury does not necessarily remove jobs from the system as soon as they are complete - a completed job may linger for a while after completion before being removed. The counters are not updated in real time, but sufficiently near to it that they'll give you a reasonable idea of the state of your queue. Their main value is as an indicator that something is wrong somewhere: if you see a huge number of pending jobs, then that may indicate that there is a connection or upstream problem of some kind. Similarly, huge numbers of Ready jobs may indicate a denial of service attack of some kind. Killed jobs end up in the "Complete" column until they're removed. "Complete" counter resets are affected by time, queue activity and the type of failure, and may vary over time. Cheers! -- David --