I am trying to understand the throttling behavior that occurs with azure search queries. We are currently running on 3 replicas and 1 partition which seems like it should be more than enough to deal with 14 requests/sec and yet we see that the ~1% queries are getting throttled. (I have attached an image below). Right now our users just see an error being returned for those cases.
Can I please get more info on why these queries might be getting throttled; Can we do something to prevent this kind of throttling; Is this to be expected and we need to implement error handling for throttled queries and implement some kind of retry mechanism. Are there any best practices to deal this issue. Thanks!
[![Azure Query Metrics][1]][1]