Thread-safe Sliding Window P99 Latency Tracker

Thread-safe Sliding Window P99 Latency Tracker Implement a thread-safe (concurrent) latency tracker that supports ingesting latency samples and querying the P99 over a recent time window. You need to implement two functions: setLatency(long timestamp, double latency) getP99Latency(long windowSize) 1) setLatency timestamp: the timestamp (integer) associated with this latency sample. latency: the latency value (double). This method may be called concurrently by multiple threads. 2) getP99Latency windowSize: size of the time window. Return the P99 (99th percentile) of the latency samples that fall within the most recent windowSize time range. This method may also be called concurrently, and concurrently with setLatency. P99 definition Sort the latencies in the window in ascending order. Let N be the number of samples in the window. Return the value at position ceil(0.99 * N) (1-indexed). Output convention If there is no data in the window, return an agreed value (e.g., 0.0 or NaN, as specified). Performance and concurrency requirements Must support high concurrency for reads/writes. Must control memory as data grows (evict samples outside the window). Example Assuming the window is [now - windowSize, now] (the exact definition of now should be clarified): Ingest: setLatency(1000, 10.0) setLatency(1005, 20.0) setLatency(1010, 30.0) Query: getP99Latency(20) computes the P99 over samples with timestamps in [now-20, now]. Sample test scenarios Empty window: query P99 with no samples. Single sample: P99 equals that sample. Multiple samples: verify the ceil(0.99*N) rule. Window eviction: old samples must be excluded. Concurrent read/write: multiple threads calling both APIs; no exceptions and consistent behavior. Example Input setLatency(1000, 10.0) getP99Latency(100) Output 10.0