Latency budgets at the edge: a quantitative framework for sub-millisecond ingestion
We present a framework for allocating latency budgets across a distributed ingestion pipeline, validated on a 2M-events/sec production workload. By moving validation and filtering to a Rust/WASM edge runtime, p99 ingestion latency fell from 41ms to 0.8ms while compute cost dropped 60%.
We present a framework for allocating latency budgets across a distributed ingestion pipeline, validated on a 2M-events/sec production workload. By moving validation and filtering to a Rust/WASM edge runtime, p99 ingestion latency fell from 41ms to 0.8ms while compute cost dropped 60%.