Hey everyone. We’ve been aware of a number of reports related to connections to the Twitter streaming over the past few days.
Thank you for the reports. First of all, there is no “everything on lock down” going on here, as was mentioned as a possibility in one thread - the Dyn issues last week clearly affected us as you’ll see from Tweets from @Support and may be related to these reports, but there’s no deliberate change that I’m aware of.
I am trying to work through 4 threads open on the forums (and some Tweets) to see whether I can identify a common and specific pattern between the reports.
As far as I understand it, the 420 error should only happen on connection - not during a connection, since the stream is already open. Several of the reports here seem to refer to very short-lived and then broken connections across public sample, user streams, and site streams. Presumably you mean that when you then try to reconnect you see the 420 error.
Creating new account and application keys are very unlikely to help in situations like this as the system will just start to log multiple connection attempts from different apps on single IPs, which is problematic (unless you are authorised for site streams).
What I believe I’m seeing in common here is:
- all of the streaming endpoints seem affected
- connections that have previously seen no issues now seem to disconnect 1-5 minutes after opening, and subsequent connection attempts produce 420 errors
- in the case of Site Streams, multiple connections are not being allowed
I’d like to keep the investigation as focussed as possible so let’s please try to limit this thread to related discussion rather than everyone saying “me too” or “oh and I have this additional issue”. Equally please do not repeat what you posted in the other 4 threads, we’ve seen them.
Please bear with us - and where possible, please help us to narrow down the focus across these reports!