API Server hangs up the socket on stream / update requests

eknkc · 2013-08-22T20:21:10.000Z

Hi,

We’ve been using stream and rest apis for 2 years now, aside from a 1.1 upgrade, not much changed and everything has been working fine.

However, since yesterday, our filter stream API requests are being dropped without any response by the servers. Socket disappears and that’s it. However, I noticed that it only happens when we send a track paramter with 400 items in it. If I send a smaller request, it passes through just fine.

Same thing happens in update_with_media rest endpoint. Since the data is normally larger than a couple hundred bytes, it gets dropped just like the stream request.

Any idea? There has not been any code changes and we strictly follow the rate limits on our calls. We open a single stream and back off during retries.

Thanks.

LogicalArthur · 2013-08-26T18:20:53.000Z

Hi Ekin:

Can you please let me know the following items of information?

Username under which you’re trying to connect (i.e., the username the app is associated with)
The app name
The IP address(es) of your servers trying to connect to the stream
The date and time ranges when you experienced the difficulty

Thank you very much.

eknkc · 2013-08-26T21:30:00.000Z

Hi Arthur,

Username: @onediocom (other user tokens can not use update_with_media as well)
App: Onedio Com (under user @onediocom)
Main IP: 188.132.223.90 (onedio.com). 31.210.61.106 also has difficulities, it’s on the same datacenter.
It has been like this since August 22 10:00 AM UTC, ongoing.

Also, it seems that half of our cards are not being parsed by the Twitter Card crawler: https://twitter.com/onediocom All tweets with a link to oned.io had it’s card available until Aug 22th. Then it became unstable.

Thanks!

eknkc · 2013-08-27T13:47:40.000Z

It seems to be working properly now. We did not restore full service yet though, however our tests seems to indicate no problems. So is the problem resolved completely now?

Thanks a lot!

MeanwhileMedia · 2014-06-07T13:21:06.000Z

@LogicalArthur, @eknkc - Was something done to resolve this issue? I am running into the same issue and would love some input.

thanks