[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"glossary-backpressure::en":3,"gloss-cluster-backpressure::en":26,"gloss-next-backpressure::en":9},{"slug":4,"category":5,"name":6,"definition":7,"meta_desc":8,"faq":9,"schema_markup":9,"related":10},"backpressure","data-infra","Backpressure","Backpressure is a system's way of signaling upstream that it can't keep up, so producers slow down instead of overwhelming a consumer. When data arrives faster than it can be processed, something has to give: without backpressure, queues grow unbounded, memory fills, and the service eventually crashes.\n\nWell-designed streaming and messaging systems make this explicit. A consumer that's saturated stops pulling new messages; a bounded queue that's full blocks or rejects new writes; a reactive stream propagates a 'slow down' signal all the way back to the source. The alternative — dropping data silently or falling over — is almost always worse.\n\nFor SaaS builders, backpressure is what keeps a traffic spike from turning into an outage. Practical note: prefer bounded queues and buffers over unbounded ones, decide up front what happens when a limit is hit (block, shed load, or route to a dead-letter queue), and load-test the failure path, not just the happy path.","Backpressure lets a consumer tell producers to slow down instead of drowning — without it queues grow unbounded, memory fills, and the service eventually falls over.",null,[11,14,17,20,23],{"slug":12,"name":13},"batch-processing","Batch Processing",{"slug":15,"name":16},"dead-letter-queue","Dead-Letter Queue (DLQ)",{"slug":18,"name":19},"delivery-semantics","Delivery Semantics",{"slug":21,"name":22},"message-queue","Message Queue",{"slug":24,"name":25},"streaming-data-processing","Streaming Data Processing",[27,31,34,35,39,42,45,48,51,55,58,61],{"slug":28,"category":5,"name":29,"updated_at":30},"acid","ACID","2026-08-24T02:46:37+00:00",{"slug":32,"category":5,"name":33,"updated_at":30},"ann-search","ANN Search",{"slug":12,"category":5,"name":13,"updated_at":30},{"slug":36,"category":5,"name":37,"updated_at":38},"bm25","BM25","2026-08-24T02:46:38+00:00",{"slug":40,"category":5,"name":41,"updated_at":30},"cache","Cache",{"slug":43,"category":5,"name":44,"updated_at":30},"cap-theorem","CAP Theorem",{"slug":46,"category":5,"name":47,"updated_at":30},"change-data-capture","Change Data Capture (CDC)",{"slug":49,"category":5,"name":50,"updated_at":30},"chroma","Chroma",{"slug":52,"category":5,"name":53,"updated_at":54},"chunk-overlap","Chunk Overlap","2026-08-24T03:30:02+00:00",{"slug":56,"category":5,"name":57,"updated_at":30},"columnar-storage","Columnar Storage",{"slug":59,"category":5,"name":60,"updated_at":30},"connection-pooling","Connection Pooling",{"slug":62,"category":5,"name":63,"updated_at":30},"cosine-similarity","Cosine Similarity"]