Re: Data Center operations mail list?

[Rafael Possamai <rafael () gav ufsc br>]

My 2 cents: I use it for other services and haven't had any issues over the
past few months, but one problem I was having with SES + Mailman is that
even though my account was out of their sandbox, I still had some smtp
errors due to "email not verified" which is annoying. So I had to tell
mailman to wrap every message, hence the via NADCOG you probably seen
before. Now that option is back to default by using Chris' server.

Their support sent me a canned message so I decided not to waste too much
time there. As long as 99% of members get their emails I don't think it
really matters whose server they are going through.

Honestly, most things out there are designed to fit the 95th percentile
scale, so if you are on either extreme, one is better off figuring out how
to adapt than to require the whole system to change, that is, if your email
server is blocking more messages than it should, fix your email server,
don't try to fix the whole world wide web.







On Fri, Aug 21, 2015 at 8:49 PM, Mike Hammett <nanog () ics-il net> wrote:

> I'm on a mailing list hosted at Amazon, uses their API, etc. Other than
> the bumps in the migration to Amazon, I haven't seen any real issues.
> Hundreds of people on the list posting hundreds (total, not each) of
> messages per day. No complaints. *shrugs*
>
>
>
>
> -----
> Mike Hammett
> Intelligent Computing Solutions
> http://www.ics-il.com
>
> ----- Original Message -----
>
> From: "Rich Kulawiec" <rsk () gsp org>
> To: "Rafael Possamai" <rafael () gav ufsc br>
> Cc: nanog () nanog org
> Sent: Friday, August 21, 2015 8:46:00 PM
> Subject: Re: Data Center operations mail list?
>
> On Fri, Aug 21, 2015 at 08:18:59PM -0500, Rafael Possamai wrote:
> > Quick update: I moved away from Amazon SES to a private smtp server
> > provided by Chris, who is also helping moderate the list.
>
> That's a good idea. I noticed.
>
> > I left Amazon SES configured as a backup since the bounce rate after
> > thousands of emails peaked at only 0.08%
>
> The bounce rate is not an effective metric, for a number of reasons, not
> the least of which is that some unknown and unknowable number of sites
> are configured to quarantine email. (This is a horrible idea that I've
> railed against many times, but that notwithstanding, ignorant people do it
> every day.) Any site which quarantines mail will not generate a bounce
> (or a reject) but will silently consign incoming traffic to a location
> which may, or may not, be eventually seen by a human being.
>
> The bounce rate yields precisely zero insight into the extent of this
> problem. Nor does it yield any insight into other similar (related)
> problems which are not manifested via the SMTP transaction.
>
> The best course here is to completely avoid any contact with the
> horribly-mismanaged Amazon cloud operation until such time as those
> running it demonstrate a bare minimum of professionalism -- which,
> to date, they have unfortunately not. In this particular case, it
> would be preferable to defer/queue any outbound mail traffic instead of
> attempting to deliver via Amazon: there is unlikely to be anything
> traversing that mailing list which would suffer by being delayed
> by an hour or a day.
>
> ---rsk