Detect Anomalous Request Patterns | Dataford Interview Questions

Write an algorithm to detect anomalous patterns in user request logs.

Treat a pattern as anomalous when the same user requests the same endpoint at least threshold times within an inclusive time window of window seconds. Implement detect_anomalies(logs, window, threshold), where logs is an unsorted list of dictionaries containing integer timestamp, string user_id, and string endpoint fields. Return each anomalous [user_id, endpoint] pair once, sorted lexicographically. Return an empty list when no pattern is anomalous.

Contract

1 <= threshold, 0 <= window; timestamps are integer seconds.

Constraints

1 <= len(logs) <= 2000
0 <= logs[i]["timestamp"] <= 10^9
logs[i]["user_id"] and logs[i]["endpoint"] are non-empty strings
0 <= window <= 10^9
1 <= threshold <= len(logs)
The same log entry may share a timestamp with other entries

Function Signature

def detect_anomalies(logs, window, threshold):

Your question is Detect Anomalous Request Patterns. Start with the requirements in the Question tab.

Run and submit as often as you like. When you're ready, talk me through your approach or go straight to the code.

Problem