Katkat
Back

Implement File Tail and Sensor Health

ConfluentSoftware Engineer · Onsite
MediumSoftware EngineerOnsiteCoding & Algorithms

Last updated

This problem tests your ability to efficiently read the end of a growing file using file pointers and to implement a sliding time-window check for sensor heartbeat events.

Part 1: File Tailing

Implement a function tail_file(filepath, n) that prints the last n lines of a file. The file may be very large (larger than memory), so you cannot load the entire file into memory. Assume the file is ASCII encoded.

Part 2: Sensor Health

You are receiving a stream of sensor heartbeats. Each heartbeat is a tuple (sensor_id, timestamp). Implement a class SensorMonitor to track the health of sensors.

  • record_heartbeat(sensor_id, timestamp): Stores the heartbeat.
  • get_unhealthy_sensors(current_time, threshold): Returns a list of sensor_ids that have not sent a heartbeat in the last threshold seconds, relative to current_time.

Example

record_heartbeat("A", 100) record_heartbeat("B", 105) get_unhealthy_sensors(120, 10) -> Returns ["A"] because B's last heartbeat (105) is within the [110, 120] window, but A's (100) is not.

Constraints

  • For Part 1: Minimize I/O operations and memory usage.
  • For Part 2: record_heartbeat should be O(1) or O(log N). The threshold is a constant integer.

Hints

Hint 1

For file tailing, use `file.seek()` to jump to the end of the file and read backwards in chunks until you have gathered enough newline characters.

Hint 2

For sensor health, use a dictionary to store the `sensor_id` as the key and its latest `timestamp` as the value.

Hint 3

Ensure you handle edge cases for the file tailing, such as files that are smaller than the requested number of lines or empty files.

Follow-up questions

  • How would you modify the file tailing solution if the file is being continuously appended to by another process?
  • If the number of sensors grows to millions, how would you optimize memory usage in the SensorMonitor?
  • What concurrency issues might arise if multiple threads call `record_heartbeat` and `get_unhealthy_sensors` simultaneously?