Mailwizz 1.3.9.x - feeling brave to test the beta release?

Status
Not open for further replies.
frm.mwz said:
And the custom-rules still overrides? Also, is a bounce overview table with categorization possible (this way, one could see the bounces and put them into action categories)? Plus, the GiveUp would be important to be able to adjust in number and when to repeat (same/next cron same/next DS, etc)?


This is very good, can you add also minute and second please.

In any case, great developments :cool:
Click to expand...

Hello @frm.mwz, I wanted to know is new feature "hourly quota for customers and customer groups" working with you? Because I'm in discussion with @twisted1919, that by adding even 10K emails system sends more than 10K emails.
 
AC AFD said:
Hello @frm.mwz, I wanted to know is new feature "hourly quota for customers and customer groups" working with you? Because I'm in discussion with @twisted1919, that by adding even 10K emails system sends more than 10K emails.
Click to expand...
If it is re paused/restarted campaigns that were there before the update, maybe, but for new campaigns created/copied, under the new version, it seems to work with customer group quotas and delivery server quotas as set, incl change server at say 1k for a whole 20k, ie 20x.
Please let me know the exact settings you used and the results you got, and we will then try to replicate.
 
frm.mwz said:
If it is re paused/restarted campaigns that were there before the update, maybe, but for new campaigns created/copied, under the new version, it seems to work with customer group quotas and delivery server quotas as set, incl change server at say 1k for a whole 20k, ie 20x.
Please let me know the exact settings you used and the results you got, and we will then try to replicate.
Click to expand...
@frm.mwz, Thanks for your reply, messaged you all the details and we're running new campaigns :)
 
@twisted1919
great update and hourly quota is nice addition but i think we are missing daily quota too.
like on Sparkpost we have time daily quota and we want to assign that daily quota to our customer or group but we can't. can this daily quota be implemented ?
 
@AHK - For now i won't be adding it because i want to see how the hourly quota will act in terms of performance, then based on that we will see if a daily quota worths the effort or not.
The daily quota can be created using the hourly quota, just divide it to 24 hours and set the result as the hourly quota for now.
 
AC AFD said:
@frm.mwz, Thanks for your reply, messaged you all the details and we're running new campaigns :)
Click to expand...

Following up on the switching/quota errors, another test with several DS set to have all the same limits and the cron to match that limit did not work in terms of switching server at (with that same limit), but it overran the quota and switched significantly later randomly.

It seems to set "giveups" as 'sent' "yes" (bug), seems to miscount them for quota usage (bug), and then continues to use the same DS even though it gets 4xx errors a couple hundred times more, and then (randomly) switches between the other available DS.

Perhaps it would be good to take care of the 4xx errors by quickly dropping (marking it as used up for the current period) the DS where they came from, categorize those correctly as 'sent' "no" and then keep them in the queue and send them at the end of the campaign again with the next cron run after the whole camp is through (I thought it did that, but could not see it), and then with another DS (if available).

Further differentiation by 4xx error might be useful.
# "421-Service unavailable" / "421 Too many connections." could be used to reduce connections, if possible (perhaps very hard to do during pcntl).
# "450-Requested mail action not taken: mailbox unavailable" / "450 Mail send limit exceeded." could be used to mark the DS as "unavailable" since the quota is used up for that period (this seems like the easiest, but could still be hard while pcntl is running).
# "451 Requested action aborted: local error in processing" could used to slow down as well, since it is oft related to DNS overload.

While the above may seem not really a mwz problem, it is in the end something the mwz user needs to cope with, as the switching needs to work under various (not just the above circumstances), and as such it would really be good to resolve this step-by-step.

Any better ideas/solutions most welcome!
Thx ;)

Might amend this post with further observations/suggestions.
 
Last edited:
@frm.mwz - mailwizz is not a load balancer, it's role is to shove as many emails as possible to a server that can cope with the load. If the server isn't able to keep up with mailwizz then you have to do find a way to do it.
We have to stop putting mailwizz do things that it isn't suppose to do.
 
twisted1919 said:
@frm.mwz - mailwizz is not a load balancer, it's role is to shove as many emails as possible to a server that can cope with the load. If the server isn't able to keep up with mailwizz then you have to do find a way to do it.
We have to stop putting mailwizz do things that it isn't suppose to do.
Click to expand...

Well, I completely agree, the above is just one way of seeing certain errors, the counting and the switching seems is sometimes not working as one would perhaps expect in these cases, but it got better!

You know I love MailWizz, and as many depend on these features, I think it would help if they are taken care of step-by-step:
# giveups sometimes not counted correctly
# quotas sometimes not counted correctly
# resending of giveups seems not to happen sometimes
# switching of delivery servers does not happen as set sometimes
# the above seems especially so during pcntl, less oft without.
Many thanks for looking into these!
:rolleyes:
 
Status
Not open for further replies.
Back
Top