Advertisement
Producer pushes; consumer pops. If producer > consumer, queue fills and blocks.
What you're seeing
A bounded queue between producer and consumer absorbs short-term rate mismatches. If producer sustained rate > consumer rate, the queue fills. Then producers block (or drop messages).
This is the queueing-theory core of every backpressure system: web servers, message brokers, stream processors, network buffers. Set queue capacity = (acceptable latency × consumer rate).
★ KEY TAKEAWAY
If producer rate > consumer rate, the queue fills. With bounded queue + blocking, producers throttle. With dropping, you lose messages.
▶ WHAT TO TRY
- Slide producer rate higher than consumer.
- Watch the queue fill, then block ratio rise.
- This is the queueing-theory model behind every backpressure system.