hardSystem Design

How would you design an analytics/metrics pipeline?

Q: How would you design an analytics/metrics pipeline?

## Analytics / metrics pipeline You're ingesting a **firehose** of events (clicks, page views, app metrics) and need both **real-time** dashboards and **accurate** historical reports. The standard shape: buffer everything in a durable **log** (Kafka), then branch into a fast **stream** path and a thorough **batch** path. Producers events Kafka event

523 views

Understand the problem

Outline a data pipeline.

analyticspipeline

Attempt it yourself

Sketch your approach before reading the solution — that's what interviews test.

Stuck? AI Nudge Available

Get a conceptual hint to guide your logic without spoiling the final implementation.

Study the solution

The solution is waiting

Give it an honest attempt first — then compare your thinking with the full walkthrough.

Read the code

Tumbling-window aggregation

Run Playground

from collections import defaultdict

# (timestamp_seconds, metric) event stream
events = [
    (0, "page_view"), (5, "page_view"), (61, "page_view"),
    (62, "click"), (119, "page_view"),
]
WINDOW = 60   # seconds per tumbling window

agg = defaultdict(int)
for ts, metric in events:
    window = ts // WINDOW                 # bucket the event into its window
    agg[(window, metric)] += 1

for (window, metric), count in sorted(agg.items()):
    print("minute", window, metric, "=", count)

Join the discussion

Discussion (0)

No responses yet. Be the first to share what you think.

How would you design a rate limiter?

How would you design a URL shortener like TinyURL?