> On Mar 24, 2016, at 11:34 PM, Aaron Routt <aaronro...@gmail.com> wrote: > > The only logs I can access is the receipt-
You'll have to do better than that if you want help. > > Mar 19 19:00:13 mail-d1f9ab60 postfix/smtp[30442]: 3qQpd33djJzDywj: > to=<REDACTED>, relay=cluster1.us.[REDACTED].com[XXX.XX.XX.XXX]:25, > conn_use=66, delay=193766, delays=193728/38/0/0.1, dsn=2.0.0, status=sent > (250 ok 1458414013 qp 36488 > server-16.tower-192.REDACTED.com!1458414005!48124713!66) You have a backlog of email to this messagelabs hosted domain. The qshape tool might also shed some light on the situation. See http://www.postfix.org/QSHAPE_README.html > > Funny thing is the delays started at ~2000+, then after a handful went up to > 36000+, then after the first 300 deliveries the delay jumps up to 180000+, > which looks like basic throttling to me. I have a hard time believing it is > my mail sever not sending the mail (though entirely possible by the ISP but > I've ruled that out), versus their mailserver putting it in delayed (you can > probably guess the service redacted above). If the input rate exceeds the output rate for long-enough a backlog forms, and "conn_use" numbers become high as Postfix engages demand caching. > > On Thu, Mar 24, 2016 at 8:14 PM, Viktor Dukhovni <postfix-us...@dukhovni.org> > wrote: > > > On Mar 24, 2016, at 10:40 PM, Aaron Routt <aaronro...@gmail.com> wrote: > > > > delays=193728/38/0/0.1 > > Insufficient context. Post all the log entries for the queue-id > of the message that logged this combination of delays. You need to analyze the your logs, there's no silver bullet. Understand how quickly mail for this domain is arriving, how quickly it is leaving, what your concurrency settings are, any artificial rate delays, ... Skimping on the data you post and/or analyze is a waste of time. -- Viktor.
