Hi everyone,
I registered a Twitter API back in 3/2022 for tweets search, my program uses Twython and works properly. But recently I discovered something odd with the runtime.
I use one API to search tweets related to specific 7600 keywords, each keyword searched 10 times. The total runtime is merely 3.5 hours, but the runtime has recently increased significantly to 6.5 hours. Since my program deployed on AWS so when I check the log on the Cloud Watch, I noticed that there is something unusual with the API rate limit. Before 2022/12/25 after around 600 keywords, the rate limit is triggered and my program was delayed 15 minutes, but since 2022/12/26 just only after around 300 keywords, the rate limit is triggered so that is why my program runtime is doubled.
For the rate limit triggered mechanism, my program uses the Twython exception class TwythonRateLimitError.
My question is why in the same program with the same source code, the behavior of the API rate limit is not the same?