Design a real-time service that ingests purchase events with fields (customer_id, item_id, timestamp) and continuously outputs the top-K most purchased items over a configurable window (e.g., last 5 minutes and last 24 hours). Specify data model, ingestion pipeline, aggregation strategy (e.g., keyed windows), algorithms/data structures for maintaining top-K at scale (e.g., heaps, count–min sketch + heap), handling late/duplicate events and out-of-order timestamps, state management and TTL, fault tolerance and exactly-once semantics, horizontal scalability, and APIs for querying current top-K.

This question evaluates proficiency in real-time stream processing, stateful aggregation and ranking, event-time windowing, and distributed systems concerns such as fault tolerance, deduplication, and scalability.

How do I approach System Design interview questions?

System Design questions require understanding of core concepts and practice. PracHub provides solutions with explanations to help you master system design interviews.

What difficulty level is this interview question?

This is a hard difficulty System Design question, commonly asked during Technical Screen rounds at Amazon.

What role is this question designed for?

This question is commonly asked for Software Engineer candidates at Amazon during technical interviews.

Design real-time top-K products service | Amazon Interview Question

System Design: Real-Time Top‑K Purchased Items Over Rolling Windows

Design a real-time service that ingests purchase events and continuously outputs the top‑K most purchased items over configurable rolling windows (for example: last 5 minutes and last 24 hours).

Input Event

Each purchase event contains:

customer_id
item_id
timestamp (event time)

Assume K and window sizes are configurable at runtime.

Requirements

Specify and justify:

Data model for events and results.
Ingestion pipeline and partitioning.
Aggregation/windowing strategy (e.g., event-time keyed windows, panes).
Algorithms/data structures to maintain top‑K at scale (e.g., heaps, Space‑Saving, Count–Min Sketch + heap), including complexity and memory trade‑offs.
Handling late, duplicate, and out‑of‑order events (watermarks, allowed lateness, deduplication).
State management, TTL/retention, and backpressure control.
Fault tolerance and exactly‑once processing semantics end‑to‑end.
Horizontal scalability and hot‑key mitigation.
APIs for querying the current top‑K for given windows (and update cadence/latency).

Make minimal, explicit assumptions as needed (e.g., target throughput, latency SLOs), and propose a design that works for both a short window (5m) and a long window (24h).

Requirements

Specify and justify:

Data model for events and results.

Ingestion pipeline and partitioning.

Aggregation/windowing strategy (e.g., event-time keyed windows, panes).

Algorithms/data structures to maintain top‑K at scale (e.g., heaps, Space‑Saving, Count–Min Sketch + heap), including complexity and memory trade‑offs.

Handling late, duplicate, and out‑of‑order events (watermarks, allowed lateness, deduplication).

State management, TTL/retention, and backpressure control.

Fault tolerance and exactly‑once processing semantics end‑to‑end.

Horizontal scalability and hot‑key mitigation.

APIs for querying the current top‑K for given windows (and update cadence/latency).

Make minimal, explicit assumptions as needed (e.g., target throughput, latency SLOs), and propose a design that works for both a short window (5m) and a long window (24h).

Design real-time top-K products service

Quick Overview

System Design: Real-Time Top‑K Purchased Items Over Rolling Windows

Input Event

Requirements

Solution

Comments (0)

Design real-time top-K products service

Quick Overview

System Design: Real-Time Top‑K Purchased Items Over Rolling Windows

Input Event

Requirements

Solution

Comments (0)