Twitter Streaming API down?

streaming

#1

I’m able to oauth to the streaming API, but I’m not getting any tweets. I know Twitter went down for a little while this morning, and is back up. Is the Streaming API up and running?

Brian Maso


#2

Hi Brian, the streaming API is currently unavailable and there are performance issues on a number of other services too, you can watch https://dev.twitter.com/overview/status for updates.


#3

Seems to be affecting SiteStreams too, get a connection OK but no content delivered back down


#4

Not sure if this is helpful info or not: I am not getting any tweets when connecting to the streaming API on a box in an AWS data center, but I am able to get tweets when connecting to the Twitter streaming API through a local network (my cell phone, actually).

I am guessing this means some streaming servers are working, and others aren’t. I am just lucky to connect to a working one through my cell phone.

Brian Maso


#5

According to the status page user streams are still up which explains what you are seeing


#6

Apologies to everyone. There was a hardware issue which affected some streaming API connections. When I hit the office I posted a status update on @twitterapi but I wasn’t able to acknowledge it earlier here myself.

Working on getting things back to normal.


#7

I think the issue affected calls to
https://api.twitter.com/1.1/application/rate_limit_status.json?resources=search

Rate limit is not being updated.


#8

Hi Andy,

Need help with this one: attempts to make site stream connections for @meshfire are failing with 420 error. At this point, we have 9 streams open, most of which have a couple hundred users on them.

Attempts to start development copies of the application are the ones getting the 420 error. Could be related to the assertion I see in this thread that rate limits are not being updated correctly?

Has anyone else seen this? Seems like our usage should be well below what the rate limits require. Thanks!

Best,
Eli Israel
eli@meshfire.com


#9

Oh, I should add: this is a new symptom that only happened after this morning’s brief outage. Thanks.


#10

Things have been returning to normal and the issue was largely resolved by about 1630 UTC. If you’re still seeing issues, I think you should have a contact email address for support with site streams questions, and the team should be able to look at the question in more detail.


#11

Thanks, Andy. We are indeed still receiving the 420 error, often preceded by a 503 when attempting to connect to the control stream.

I’ve sent a couple of emails to the contacts I have within the site streams team, but honestly things have been so solid, it’s been a while since I had to reach out to them. Would you do me a favor and ping them as well so that I can make sure I have the right person as soon as possible.

PS, will we see you at Flight?


#12

Quick followup: We appear to have cleared the error by shutting down and restarting all our streams. Somewhere in the mix, connections that seem to have been hung up from before the outage were perhaps causing confusion. Thanks, Andy, we’re good now.


#13

Good to hear - see you at Flight!


#14