ClickHouse adds Silk fiber runtime to boost query scheduling

ClickHouse announced Silk, a new fiber runtime that reworks query scheduling and execution [ClickHouse Blog]. Silk replaces the previous scheduler with a lightweight fiber‑based engine, aiming to cut query latency and raise throughput for workloads such as real‑time analytics and large‑scale data processing [ClickHouse Blog].

Silk integrates directly into ClickHouse’s execution pipeline. By keeping query tasks on a single OS thread and switching between them via fibers, the runtime eliminates context‑switch overhead and allows the server to keep more work in‑flight. The blog notes that this design should lower latency and increase throughput compared with the prior implementation.

Because Silk is open‑source and shipped under the Apache 2.0 license, the community can inspect, modify, and extend the codebase. The addition also gives engineers flexibility in deployment: resource allocation, query prioritisation, and configuration can be tuned without altering the core server binary, according to the announcement [ClickHouse Blog].

Silk’s introduction is the latest step in ClickHouse’s effort to optimise its query engine for high‑performance use cases. By focusing on the scheduler, the project targets the part of the stack that often becomes a bottleneck in large‑scale analytics workloads.