Limpio – New Crypto Market Data Engine in Go (Redis and Timescale)

The docs describe a sensible hot/cold architecture — Redis + WS for the real-time hot path and TimescaleDB as a cold store — with event-driven NEW_CANDLE triggers, batched indicator computation and worker pools to control tail latency. It reads like an engineer's playbook for predictable p95/p99 behavior, but the writeup is mostly architectural: I'd want hard benchmarks, deployment artifacts or source to be convinced this outperforms established market-data stacks in practice.