No more than 10 simultaneous connections allowed.

[Michael D. Yaeger]

We have been using the MailChimp API to create and send newsletters
from our website for a couple of years. However, we have begun to have
issues in connections. We have verified that we were not using the
deprecated login() function noted in the links MC staff sent:
http://eepurl.com/fpsn2
http://eepurl.com/g-h3I

The error we are receiving is "No more than 10 simultaneous
connections allowed." Note that this error only began to occur around
the end of October or first of November. It has plagued us ever since.
The only way to send a newsletter is to wait about 1.5 hours after
receiving the error and try again.

We are on a VPS with dedicated IP address. There is usually only 1
person working on the newsletters at a time.

[jesse]

Those have been in place for quite some time:

http://apidocs.mailchimp.com/api/faq/#faq6

You'll need to take a look at your code and/or your API dashboard to
determine why so many calls are being made on top of each other:

http://apidocs.mailchimp.com/api/faq/#faq3

You may also want to review this list to see what you may be doing
that is totally unnecessary:

http://apidocs.mailchimp.com/api/faq/#faq5

jesse

On Dec 14, 2:54 pm, "Michael D. Yaeger" <myae...@k3tech.com> wrote:
> We have been using the MailChimp API to create and send newsletters
> from our website for a couple of years. However, we have begun to have
> issues in connections. We have verified that we were not using the

> deprecated login() function noted in the links MC staff sent:http://eepurl.com/fpsn2http://eepurl.com/g-h3I

[wayan]

Since yesterday we (mailchimp account bonybony) encounter the same
error : No more than 10 simultaneous connections allowed.
We call the API (XML RPC) from 217.31.59.112, there is one daemon, one
thread. If I stop the daemon and try api call from command line, the
result is same.

Our sysadmin explicitly searched the netstat for connections to
mailchimp. None.

I tried to get some clue in the account, but "Your account is
currently under maintenance. Please try again later" for 16 hours at
least.

Where are those 10 connections come from? Why such long maintenance?

Roman Daniel

[jesse]

We tracked that down - when the "under maintenance" message was being
returned connections were being improperly counted, so basically just
the wrong error message was returned - either way the calls wouldn't
have worked. I can also once the maintenance was completed on your
account it stopped being triggered. We have a fix ready to go out to
make sure the proper error is returned in the future.

jesse

[Michael D. Yaeger]

I sent a reply yesterday, but it didn't post...

Is the limit 10c/s, 10c/m, 10c/h, or what? Checking the API dashboard,
there has been no time when we ever hit 10 SIMULTANEOUS connections.
The most connections we have in a day are 30-40, ALL DAY. When we send
a newsletter there can be 10 connections within in a minute, but that
is because we make edits, then click save. We do that a few times to
make sure we don't loose work. Other than that, everything looks fine.

As a side note, none of the dates/times are correct in the dashboard.
It is currently showing connections 5 hours into the future.

[jesse]

All limits are in the first link in my first reply, and simultaneous
means simultaneous. If you got caught up in the maintenance stuff, my
last reply may be valid. You may also want to double check the
account(s) being used that are actually getting the error if you are
accessing more than one.

All API related stuff is always in GMT.

jesse

[Michael D. Yaeger]

The link states "10 simultaneous connections per IP address (yours)
per user account" which would mean 10 open connections at any given
time, regardless of how far apart they are made. There is no way this
is happening. We were never getting the error before, and since
getting the error, we have added explicit "close connection" calls.
Since these connections are being made from a single PC, this seems
impossible to accomplish!

In regards to the time, could you add a feature request that syncs the
API time with time zone in user preferences? I don't know anyone who
uses GMT.

[Michael D. Yaeger]

Jesse, the only account I am referring to is "steelmarketupdate" which
is the only account on this server/IP.

On Dec 16, 10:46 am, jesse <je...@mailchimp.com> wrote:

[Michael D. Yaeger]

Also, can you explain the green "check" and red "stop" in the
"Results" column. There was no documentation or title/comments I could
find.

On Dec 16, 10:46 am, jesse <je...@mailchimp.com> wrote:

[jesse]

There's nothing to this beyond what I already mentioned on the 15th:

http://groups.google.com/group/mailchimp-api-discuss/browse_thread/thread/7a42555a0a9134fe#484b12cb584097f7

Digging through our logs, you received this error:
2011-12-08: 6 times
2011-12-09: 14 times
2011-12-14: 4 times

And nothing since. So you really have been drastically overstating any
effect or issue receiving that error caused - after that, you should
handle errors and retry calls later as necessary.

Results:
Green Check marks == good == call returned successfully

Red Stop signs == bad == call returned an error


jesse

[Michael D. Yaeger]

Jesse, I thought that response was to Wayan as we were not getting an
"under maintenance" message in our account. We have been using a
workaround ever since, as we were still waiting on a response to our
issue. We will test normal publishing procedures tonight and let you
know if the "10 connection" error still exists.

In regards to the stop sign errors, these seem to be coming from a
chron job which removes our expired members and adds new ones. I
suppose a specific message is being returned. Is there anyway we can
see those errors from previous dates or can we only log new ones?

On Dec 22, 11:21 am, jesse <je...@mailchimp.com> wrote:
> There's nothing to this beyond what I already mentioned on the 15th:
>

> http://groups.google.com/group/mailchimp-api-discuss/browse_thread/th...

[jesse]

You'll need to log errors if you want to be able to investigate them -
what you see is the extent of the detail we store. Also, your call
patterns indicate that nothing has stopped, so you may want to get a
better handle on what you're actually looking at.


jesse

[Dustin Moskovitz]

Jesse, we are also encountering this error in a script I'm working on. Similar to Michael, I am confident we do not actually have 10 concurrent connections. Can you say anything about how that is determined on Mailchimp's side?

[jesse]

This explains it:
http://apidocs.mailchimp.com/api/faq/#faq6


jesse

[Dustin Moskovitz]

I've read that doc, but again I (like Michael) don't believe that I have 10 concurrent connections. Can you say how that is tracked on Mailchimp's side? Is it definitely the case that the tally of open connections goes down as soon as a REST call completes, or is it possible that there is a delay?

[jesse]

Yes, it is, and that faq explains how it is tracked.


jesse