We’re currently experiencing very long processing times for asynchronous analytics jobs, even for very simple calls. The jobs are stuck in the PROCESSING state. Are the performance issues described in this announcement back?
Do you have any specifics in terms of timing when this issue occurred on your end? Additionally, is this a recurring issue, or has it been resolved now?
Thanks for getting back!
The performance started deteriorating around Sep 6th and hasn’t returned to normal after that, i.e. the processing times are still much, much longer than usual. However, we found an issue on our side today which may have affected the API requests and response handling. Let’s see if fixing this issue resolves the problems with slow requests.
Sounds good! Please keep us updated.
It seems that some of the performance issues were due to an error on our side, resulting in some of the API requests being rejected. But we’re still seeing quite long processing times even for simple async jobs. Any ongoing issues on your side?
Others seem to experience similar issues. If there’s any information we can provide or anything else we can do to troubleshoot, please let us know.
We ran into these exact issues in July right after Twitter released the v2 endpoints. Our thinking is they have cut the number of resources dedicated to v1 endpoints. We’ve now migrated over to v2 and are no longer having these issues.
Hopefully this helps!
Thanks for the tip, @saraj345! Unfortunately both the v1 and the v2 endpoints seem to be equally slow for us at the moment.
For example, a simple API request that returns results in less than one second using the synchronous endpoint can take 5-20 minutes to finish using the asynchronous endpoint.