We use the twitter enterprise firehose to actively monitor relevant conversations for our clients, and one question that frequently arises has to do with twitter threads - i.e. when one of our clients is mentioned, we would like the ability to pull down the entire twitter thread.
The conversation_id appears to solve this problem, and our approach is as follows:
- We receive a relevant tweet via the firehose
- In real-time, we look up the conversation_id via the /tweets endpoint in the new v2 API.
- In real-time, we send a query: “conversation_id:{conversation_id}” using the recent search endpoint in the new v2 API.
Step #2 works correctly, as we get back a conversation_id, but step number 3 fails to return results in real-time. We do, however, get results back when we retry the same query several minutes later.
Unfortunately, this delay hinders our ability to actively monitor events in real-time.
Is there any documentation regarding latency of the conversation_id lookups via the recent search endpoint? Is there a plan to reduce the latency of these lookups via the “recent-search” endpoint? Is there a more direct way of getting this information, as opposed to using the approach described above?
Thanks!
cc: @andypiper @jessicagarson
1 Like
