perf(clickhouse): enable LZ4 client compression on inserts#911
Merged
Conversation
Closes #909 Local bench (loopback CH, 500k rows × ~700B JSON): - wall-clock: 3,223ms → 1,130ms (−65%) - wire bytes: 408MB → 39MB (10× smaller) - throughput: 155k → 442k rows/s LZ4 is the clickhouse-go maintainers' default recommendation — fast compress/decompress, negligible CPU. ZSTD compresses ~3× better but its higher CPU cost erases the wall-clock win on the native protocol; LZ4 is the right pick for event-shaped JSON payloads. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
alexluong
added a commit
that referenced
this pull request
May 24, 2026
Reproduces the LZ4-vs-ZSTD-vs-off comparison from #911 against any CH instance. Self-contained — creates and drops its own throwaway table. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Reproduces the LZ4-vs-ZSTD-vs-off comparison from #911 against any CH instance. Self-contained — creates and drops its own throwaway table. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
alexluong
force-pushed
the
fix/clickhouse-lz4-compression
branch
from
1d2f99c to
edf9ad6
Compare
alexbouchardd
approved these changes
May 24, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Closes #909
Summary
Compression: &clickhouse.Compression{Method: clickhouse.CompressionLZ4}ininternal/clickhouse/clickhouse.go.cmd/bench/clickhouse-compression/used to validate the choice (reusable if we ever revisit).Why LZ4 (not ZSTD)
LZ4 is the clickhouse-go maintainers' standard recommendation for the native protocol: very low CPU, decent ratio, fast decompress. ZSTD compresses ~3× better but costs more CPU and erodes the wall-clock win on event-shaped JSON payloads.
Bench numbers
Measured against a local CH instance (
clickhouse-server:24-alpine, dev deps stack). 500k rows × ~700B JSON each, median of 3 runs, alternating off/lz4/zstd to spread transient blips:LZ4 holds a steady ~65% wall-clock reduction across batch sizes (10k → 500k rows). Even on loopback (zero network cost) the syscall/buffer savings dominate the compression CPU cost.
Wire bytes pulled from
system.query_log.ProfileEvents['NetworkReceiveBytes']via taggedquery_id. Seecmd/bench/clickhouse-compression/README.mdto reproduce.Test plan
go build ./internal/clickhouse/clickhouse-server:24-alpine(dev deps stack)