Katkat
Back

Design a user behavior monitoring system

RipplingSoftware Engineer · Onsite
MediumSoftware EngineerOnsiteSystem Design

Last updated

This question evaluates a candidate's ability to design scalable, low-latency user event ingestion and analytics pipelines while reasoning about data storage choices and aggregate correctness trade-offs.

System Design: User Behavior Monitoring System

You are asked to design a monitoring/analytics system that collects user behavior events from clients (e.g., app install, user clicks, etc.). The system must support:

Functional requirements

  1. Real-time monitoring: show near-real-time metrics on dashboards (e.g., installs per minute, click-through rate).
  2. Raw event retention: store raw events for large offline/batch jobs (e.g., daily/weekly aggregation, deeper analysis).
  3. Metric discovery/confirmation: you should clarify with the interviewer which exact product metrics are needed (assume multiple counters/ratios, grouped by dimensions like platform/app version/country).

Non-functional requirements (assume typical expectations)

  • Handle high write throughput (many clients emitting events).
  • Low-latency updates for dashboards (seconds-level is acceptable unless otherwise specified).
  • Scalable storage and compute.
  • Reasonable correctness for aggregates (be explicit about tradeoffs: at-least-once vs exactly-once, late events).

Out of scope

  • System/infra observability metrics like host reliability/latency are not required unless you want to briefly mention them.

Design the architecture, key components, data flow, storage choices, and how you compute/serve real-time metrics vs offline analytics.