Question 1

What is a system architecture bottleneck?

Accepted Answer

A system architecture bottleneck is any component — server, database, cache, or queue — that receives more traffic than it can handle, causing slowdowns or errors for the entire system. Common examples are database bottlenecks under high read load and single-instance servers with insufficient QPS capacity.

Question 2

How do I find bottlenecks in my system architecture?

Accepted Answer

Use Bottlenecker: add a load generator and your system components (servers, databases, caches, queues) to the canvas, connect them with edges, and set the incoming QPS and each component's capacity. The simulation re-runs on every change and overloaded nodes turn red, showing where your architecture bottleneck is before it reaches production.

Question 3

How can I avoid architecture bottlenecks?

Accepted Answer

Common strategies to avoid architecture bottlenecks include: horizontal scaling (adding server instances), adding database read replicas, introducing a caching layer (Redis/CDN) to offload reads, and using message queues to buffer traffic spikes. In Bottlenecker you can try adding server instances, read replicas or a cache on the canvas and see how utilisation and error rates change before you build it.

Question 4

What causes a database bottleneck?

Accepted Answer

Database bottlenecks occur when query volume exceeds the database's read or write QPS capacity. Retry storms (where upstream servers retry failed requests) can multiply load significantly. Adding read replicas, caching frequent queries, and separating read/write paths are the most effective fixes.