Link Aggregator / Tech News

Interview-Ready Approach

1) Clarify Scope and SLOs

• •Problem statement: Design a Hacker News / Reddit-style link aggregator with voting, ranking, comments, and anti-manipulation measures.
• •Design for a peak load target around 1,157 RPS (including burst headroom).
• •Daily active users: ~500,000
• •Page views/day: ~20,000,000
• •Submissions/day: ~10,000
• •Comments/day: ~200,000
• •Front page load time: < 200 ms

2) Capacity Planning Method

• •Convert traffic and growth constraints into request rate, storage growth, and concurrency budgets.
• •Keep at least 2-3x safety margin per tier (ingress, compute, storage, async workers).
• •Reserve explicit latency budgets per hop so p95 can be defended in review.

3) Architecture Decisions

• •Databases: Define a clear system-of-record and design read/write paths separately before adding optimizations.
• •Caching: Put cache on hot read paths first and pick cache-aside or write-through explicitly.
• •API Design: Standardize API boundaries, idempotency keys, pagination, and error contracts first.
• •Search: Use primary store for writes and async index updates for search relevance + scale.
• •Rate Limiting: Enforce token/sliding-window limits at ingress and for sensitive internal APIs.

4) Reliability and Failure Strategy

• •Use strong write constraints (transactions or conditional writes) and explicit backup/restore strategy.
• •Bound staleness with TTL + invalidation hooks for critical entities.
• •Apply strict input validation and backward-compatible versioning.
• •Track indexing lag and support reindex from source of truth.
• •Return deterministic 429 behavior with clear retry headers.

5) Validation Plan

• •Run one peak-load test, one dependency-degradation test, and one failover test.
• •Verify idempotency for all retried writes and async consumers.
• •Track user-facing SLOs first: p95 latency, error rate, and successful throughput.

6) Trade-offs to Call Out in Interviews

• •Databases: SQL gives stronger transactional guarantees; NoSQL often gives better write scaling and flexibility.
• •Caching: Higher hit rate cuts latency/cost, but stale data and invalidation bugs become primary risks.
• •API Design: Rich APIs improve developer speed but can create long-term compatibility burden.
• •Search: Search index gives rich querying but introduces eventual consistency and index ops overhead.
• •Rate Limiting: Aggressive limits protect the system but can hurt legitimate burst traffic.

Practical Notes

• •Cache the ranked front page in Redis - recompute every 30-60 seconds. Individual posts and comment trees can also be cached with TTL.
• •The ranking formula requires a periodic batch job (every minute) that re-scores top posts. Only re-score posts from the last 48 hours.
• •Store comments as a tree (parent_id reference). Flatten and nest on read. For large threads, paginate children of each comment.