the extra round thing is well documented. mixed feedback at snowflake almost always means a third coding round not a decision

the O(1) push from the interviewer is frustrating when you're right that it's not achievable. some interviewers have a solution in mind and don't deviate

that event tracking problem is basically the same as LC 715 / 352 range query structure

good luck on the third round. they're usually easier once you've passed two already