</> MAANG.io
system design · 301

Advanced

Master system design interviews with scalable architecture patterns, distributed systems, and real-world design challenges.

0/480 solved 0% complete

Section 11: Doors Left Unlocked - Follow-up Feature Evolution

Executive Summary: With the design defended, the interviewer reads out the product backlog: private tags, synonyms, smart suggestions, subscriptions and automation, tag filters in work-item queries, and trending tags. Each request gets the same three beats: what it adds to the tables, the API and the events; what stays exactly as it is; and which leak channel it opens and how the existing rules close it. Almost everything lands as either one new column or table, or one new consumer of tag-events. Two requests reach flip triggers we named earlier, and we walk through them on purpose. The assignment table, the single-shard transaction and the rule "the index narrows, the owner decides" all come through unchanged.

The interviewer relaxes and reads from a list.

Interviewer: "Customers love it. Here's what product wants next. For each one, tell me what changes and what doesn't."

The trap has two sides: a design that must be torn down, or one that quietly built all of this already. The strong answer walks each feature through a door that Sections 5 and 6 left unlocked. There is one extra test that's specific to this system: every new feature is a new leak channel. If a feature shows a name, a count or a notification, it has to pass through the same funnel as the tag page.

The Evolution Map

Start by drawing where the requests attach. Most of them hang off the event stream we already publish.

Evolution map: tag shards and the outbox relay publish to Kafka tag-events, which feeds the existing product search indexers plus new consumers: a subscription notifier that hydrates via the Permission Adapter, an automation runner acting through product APIs as the rule owner, a trending counter and an offline suggestion and merge-candidate builder that feed the tag API's popular and suggest endpoints

Sign in to continue reading

The rest of this lesson is available with a free account. Signing in with Google or Microsoft is free.

Sign in to read the full lesson

Sign in to MAANG.io

Use your Google or Microsoft account — no password to remember.

Continue with Google Continue with Microsoft

Please accept the terms above to continue.