[Greylist-users] Some real world figures from a small domain

[martin dempsey]

> Initially, I thought the same thing, so I ran that scenario past Evan a
> few weeks ago.  He showed me the way to enlightenment. (bows head, gong
> sounds).

Good. It seems to be working. Even the Mirapoint server eventually got the 
messages through. Here are some real world figures from a small domain I am 
using for testing (about 20 users). Figures are for the last 12 hours (and 
the first twelve hours of greylisting):

----------------------

mysql> select count(*),sum(blocked_count), sum(passed_count) from relaytofrom;
+----------+--------------------+-------------------+
| count(*) | sum(blocked_count) | sum(passed_count) |
+----------+--------------------+-------------------+
|      147 |                505 |                83 |
+----------+--------------------+-------------------+

mysql> select count(*),passed_count,sum(passed_count) from relaytofrom group 
by passed_count;
+----------+--------------+-------------------+
| count(*) | passed_count | sum(passed_count) |
+----------+--------------+-------------------+
|       89 |            0 |                 0 |
|       50 |            1 |                50 |
|        4 |            2 |                 8 |
|        3 |            3 |                 9 |
|        1 |           16 |                16 |
+----------+--------------+-------------------+

--------------------------------------

So about 89 messages  were blocked completely (so far, a few of them are 
probably in progress and will succeed in the next couple hours). And 83 
messages were delivered. Of the 83 delivered messages, 58 were delayed by 
greylisting. Of course this is the first 12 hours so the percentage of 
messages delayed is going to be very high. By definition, the percentage 
delayed will always be 100% for the first hour of greylisting. It will drop 
from here over the next month as the database grows.