I was testing Java code using twitter4j. Using app only auth, things were runnings smoothly all day 450/15m, then I cloned the program to test some alternatives and the clone would run into 180/15m limits (and stop the program) even though I was still reading in 450 remaining limits.
This sounds strange - should not happen if you are still using the same auth method of app-only, against the same endpoints. What is the X-Rate-Limit header saying in both cases?
I figured it out, thanks.
While the tweet query limits are 450/15m, the rate limit requests were 180/15m, so I couldn’t use a loop that was constantly checking how many queries I had left every time I queried for tweets.