Which recommendation avoids rushing to add a queue and suggests alternative optimizations first?

Which recommendation avoids rushing to add a queue and suggests alternative optimizations first?

• A. Increase CPU and storage without verifying bottlenecks.
• B. Batch writes, schema optimization, and connection pooling.
• C. Add more queues with more topics.
• D. Switch to a different messaging system immediately.

Answer: (B)

This question tests optimizing performance by addressing bottlenecks before expanding queuing infrastructure. The recommended approach focuses on making existing operations more efficient rather than adding more components. Batch writes reduce per-message overhead by sending multiple messages together, which lowers serialization and network costs and increases throughput. Schema optimization cuts the size and complexity of data being stored or transmitted, reducing CPU, memory, and I/O load. Connection pooling keeps a limited set of reusable connections alive, avoiding the expensive overhead of establishing new connections for every operation and improving overall latency and throughput. Together, these practices often unlock meaningful performance gains with less architectural disruption than adding new queues. Other options tend to introduce more resources or risk without addressing the actual performance bottlenecks: adding CPU/storage without confirming bottlenecks can waste resources; increasing the number of queues adds complexity and potential coordination overhead without solving root causes; and switching messaging systems immediately can be disruptive and may not provide a faster or safer path to performance.

This question tests optimizing performance by addressing bottlenecks before expanding queuing infrastructure. The recommended approach focuses on making existing operations more efficient rather than adding more components.

Batch writes reduce per-message overhead by sending multiple messages together, which lowers serialization and network costs and increases throughput. Schema optimization cuts the size and complexity of data being stored or transmitted, reducing CPU, memory, and I/O load. Connection pooling keeps a limited set of reusable connections alive, avoiding the expensive overhead of establishing new connections for every operation and improving overall latency and throughput. Together, these practices often unlock meaningful performance gains with less architectural disruption than adding new queues.

Other options tend to introduce more resources or risk without addressing the actual performance bottlenecks: adding CPU/storage without confirming bottlenecks can waste resources; increasing the number of queues adds complexity and potential coordination overhead without solving root causes; and switching messaging systems immediately can be disruptive and may not provide a faster or safer path to performance.