Read Replicas

One-Liner

Read replicas improve read scalability at the cost of freshness.

What It Is

A database setup where writes go to a primary node and reads are served from replicas that asynchronously replicate data.

Why It Exists

• Scale read throughput
• Isolate workloads

How It Works

Writes propagate asynchronously from primary to replicas.

Tradeoffs

Pros

• Horizontal read scaling
• Lower load on primary

Cons

• Replication lag
• Inconsistent reads

Failure Modes

• Serving stale reads after writes
• Replica lag during spikes
• Misrouting critical reads

Interview Traps

• Assuming replicas are up-to-date
• Using replicas for read-after-write flows

Real-World Usage

Common for analytics, timelines, and eventually consistent reads.

Anti-Patterns

• Strong consistency assumptions
• Hiding lag from application logic

Related Concepts

• Consistency models
• Read-your-writes