How to Answer: "How do you approach database scaling under high traffic?"

“My approach to database scaling follows a progressive, evidence-based strategy. First, before adding hardware, I optimize query execution—tuning missing indexes, eliminating N+1 queries, and configuring connection poolers like PgBouncer. Second, I introduce a caching layer using Redis with cache-aside patterns for read-heavy operations, achieving 80%+ cache hit ratios. Third, I scale reads by introducing read-replicas behind a load balancer. When write capacity becomes the bottleneck, I evaluate functional database splitting (microservices) or horizontal sharding based on shard keys like tenant_id. On our analytics platform, this approach reduced primary DB utilization from 90% to 25% under 100k QPS.”

“I first check query performance using EXPLAIN plans and ensure proper composite indexes exist. Next, I implement Redis caching for frequent queries. If write volume remains high, I work with the team to set up read replicas and optimize connection pooling.”

“I make sure queries are written efficiently without SELECT * and that foreign keys are indexed. I also utilize caching so the app doesn't hit the database for static data.”