easySOCDetection
Failed Login IP Detector
Scan login events and flag the IP addresses with too many failed attempts, the first rule every SOC writes to catch brute-force attacks.
JavaScript or Python
A SOC wants a simple brute-force detection. You get a list of login events, each a pair [ip, success] where success is true or false, and a threshold. Return the IP addresses with at least threshold failed logins, sorted in ascending string order. Successful logins don't count toward the total.
Return an empty list when no IP reaches the threshold.
Examples
Example 1
Input: events = [["10.0.0.5", false], ["10.0.0.5", false], ["10.0.0.9", true]], threshold = 2 Output: ["10.0.0.5"] Why: Two failures from 10.0.0.5; 10.0.0.9 succeeded
Example 2
Input: events = [["1.1.1.1", false]], threshold = 3 Output: [] Why: One failure is below the threshold
Constraints
0 <= events.length <= 10^51 <= threshold <= 1000- IP addresses are IPv4 strings
Hints
Hint 1
Count failures per IP in a dictionary or map.
Hint 2
Filter by the threshold, then sort the IPs as strings.
Solution
Show the solution (try it yourself first)
One pass counts failures per IP, then you keep the IPs at or above the threshold and sort them so the output is stable for analysts and tests.
def suspicious_ips(events, threshold):
fails = {}
for ip, ok in events:
if not ok:
fails[ip] = fails.get(ip, 0) + 1
return sorted(ip for ip, n in fails.items() if n >= threshold)Complexity
Time O(n + k log k) for n events and k distinct failing IPs, space O(k). Real detections add a time window (for example 10 failures in 5 minutes), look at failures per account as well as per IP to catch password spraying from many addresses, and check whether a success follows the failures, which turns noise into an incident. That follow-up is what interviewers usually ask next.
Found a mistake in this problem, its tests or the solution?