But then you start seeing how your "not pre­pro­ce­ss­e­d" queue star­ts gro­win­g, and gro­win­g...

This can al­so mean things like cla­mav or spa­ma­ssas­si­n, whi­ch need to che­ck the mail be­fo­re it ge­ts queued are not kee­ping up wi­th the mail flo­w, or ma­y­be so­me IO per­for­ma­ce is­sue.

But what can you do ri­gh now to fix it?

We­ll, you can di­sa­ble spa­ma­ssas­si­n, or, in ex­tre­me ca­ses, shu­tdo­wn SM­TP so the sys­tem has a chan­ce to ca­tch its brea­th so to speak.

Of cour­se, clo­sing SM­TP means your own users can't send email ei­the­r, whi­ch su­cks.

Now the­re is a li­gh­ter al­ter­na­ti­ve: shu­tdo­wn SM­TP for tho­se who are not your user­s.

He­re's the tri­vial co­de, im­ple­men­ted as a SPP plu­gi­n:

#!/bin/dash

if [ -f /var/qmail/control/overloaded ]
then
      if [ -z "$SMTPAUTHUSER" ]
      then
              echo R451 Temporary Failure: Server overload
              echo overload: $PPID Temporary Failure: Server overload >&2
      fi
fi

And if you are da­ring and want to make your sys­tem se­l­f-­co­rrec­tin­g, ma­y­be you should cron so­me­thing like this:

* * * * * if [ `qmail-qstat  | tail -1 | cut -d: -f2` -gt 100 ];\
then touch /var/qmail/control/overloaded ;\
else rm -f /var/qmail/control/overloaded; fi

I wi­ll pro­ba­bly co­de it again in C and make it part of ra/­plu­gin­s.