mirror of
https://github.com/GreptimeTeam/greptimedb.git
synced 2026-09-28 08:05:40 +00:00
d45f5d6eaf598d9a2f90405adfc3d17c91dd67bc
14
Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
  discord9andDennis Zhuang
|
528ceb7733 |
perf(promql): reuse sliding min and max candidates (#9099)
* perf(promql): reuse sliding min and max candidates Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * test(promql): simplify extrema benchmark parameters Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * test(promql): record baseline sliding extrema SQL results Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * perf(promql): rescan windows that barely overlap Reusing candidates loses to a plain scan when consecutive windows overlap little: the deque bookkeeping then costs more than the rescan it replaces. A local Criterion run on 4096 samples at width 240 / step 240 measured 10.79 -> 22.14 us for min and 12.83 -> 19.85 us for max. Pick the evaluator once per batch from the first two windows. RangeManipulate emits one window length and one step per batch, so that sample decides for all of them, and both evaluators return identical bits, so a wrong pick costs time only. Batches that do not qualify fold each window on its own. Move the incremental state into SlidingExtrema so tests can drive it directly: the exhaustive four-sample differential test cannot reach it through a UDF call, because such a batch never qualifies for reuse. Signed-off-by: Dennis Zhuang <xzhuang@greptime.com> * fix(promql): select the extrema evaluator from batch averages Reading the window shape off the first two windows misreads the batch. RangeManipulate starts a series at max(query start, first aligned sample), so a series that begins inside the query range gets a first window covering roughly one step, and a window covering no sample at all is emitted as (0, 0). Either one closed the gate for the whole batch, including the one-hour window at a 15s step that candidate reuse was written for. Compare the batch averages instead: at least 32 samples per window, and a step advancing at most a quarter of that. Uniform batches select exactly as before, so the thresholds keep the meaning they were measured with. The 32-sample rule had also moved most of the benchmark and query-regression shapes onto the rescan, including the case built to measure reset and rebuild. Widen those windows to 40 samples, add a step at the selection boundary, and add an end-to-end case with 40-sample windows advancing 5. Signed-off-by: Dennis Zhuang <xzhuang@greptime.com> * fix(promql): ignore empty windows when measuring batch advance The advance was read from the first and last window offsets, but a window covering no sample is emitted as (0, 0). A query whose last evaluation lands exactly one window past the last sample ends on such a window, and its zero offset made a batch of disjoint windows look like one that never moved, which selected the evaluator built for overlap. Results stayed correct; the cost was deque bookkeeping on the shape the scan fallback exists for. Take the offset span over the windows that cover a sample. Empty windows stay in the window count, where they only make both conditions stricter. Signed-off-by: Dennis Zhuang <xzhuang@greptime.com> --------- Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> Signed-off-by: Dennis Zhuang <xzhuang@greptime.com> Co-authored-by: Dennis Zhuang <xzhuang@greptime.com> |
||
|
discord9
|
ba0f7acd93 |
feat(mito2): add opt-in byte-stream-split encoding for float SST fields (#9069)
* feat(mito2): add opt-in byte stream split encoding Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * fix(mito2): correct float encoding checks Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * test(compat): cover float SST encoding Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * fix(mito2): compile float encoding tests Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * fix(mito2): release parquet test writer Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * fix(mito2): register float test primary key Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * test(mito2): verify BSS write lifecycles Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * test(metric-engine): verify BSS physical SST Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * test(mito2): verify bulk BSS lifecycle Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * fix(mito2): compile bulk BSS test Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * refactor(mito2): narrow bulk encoding constructors Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * test(compat): accept generated float upgrade output Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * test(compat): accept generated float downgrade output Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * refactor(mito2): narrow bulk encoding builder Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * test(perf): add default versus BSS storage comparison Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * test(perf): align BSS reader benchmarks with prior study Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * fix(perf): parse current read benchmark averages Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * fix(perf): retain default float encoding in direct SST fixtures Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * test(perf): isolate BSS user SSTs and benchmark every file Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * docs(perf): record measured BSS storage and reader tradeoffs Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * docs(perf): expose warm scan variability and evidence limits Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * docs(perf): clarify BSS baseline and storage measurement scope Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * test(perf): model bounded mixed integer and fractional metric series Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * docs(perf): report bounded mixed BSS measurements and query regressions Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * docs(perf): qualify timings affected by concurrent host builds Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * test(perf): include float BSS comparison in default regression cases Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * fix(perf): omit unsupported float encoding option from baseline setup Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> --------- Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> |
||
|
discord9
|
555485c40e |
perf(promql): avoid per-window allocations in simple range functions (#9104)
* perf(promql): avoid per-window allocations in simple range functions Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * perf(promql): avoid copying smoothing window values Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * test(perf): cover smoothing copy removal across window layouts Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * fix(promql): skip null samples in simple range functions Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> --------- Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> |
||
|
discord9
|
97648525cf |
perf(mito2): skip proven all-match prefilters (#9066)
* perf(mito2): skip proven all-match prefilters Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * test(perf): add manual all-match prefilter reproduction Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * test(mito2): cover all-match prefilter execution paths Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * test(mito2): match prefilter fixture to sparse SST schema Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * chore(mito2): address all-match prefilter lint findings Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * test(perf): cover all-match prefilters in default regressions Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> --------- Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> |
||
|
discord9
|
7cf84892d2 |
perf(table): filter decoded rows with dynamic predicates (#9004)
* perf(table): filter decoded rows with dynamic predicates Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * fix(table): preserve unknown dynamic filter rows Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * refactor(table): use null guards for dynamic filters Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * fix(table): preserve null inputs during dynamic pruning Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * test(perf): cover frontend join dynamic filter transfer Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * fix(table): reset dynamic filters and scanner pruning state Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> --------- Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> |
||
|
discord9
|
da5cb1a190 |
perf(promql): push down last row for instant queries (#9034)
* perf(promql): push down last row for instant queries Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * test: guard instant last row correctness Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * test: update instant query explain results Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * fix: apply last row after source deduplication Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * fix: scope post-merge last row selection Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * test(perf): cover instant PromQL last row Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * fix(perf): sort generated SST rows before writing Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * test(promql): cover instant last row selection in sqlness Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * fix(promql): avoid last row hints for lossy timestamp casts Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * test(promql): preserve stale marker semantics across flushes Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * test(promql): decode dictionary labels in stale regression Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * test(promql): avoid reserved column name in stale fixture Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * test(promql): exercise LastRow hints and filtered results Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * refactor: keep after-merge mode in LastRow selector Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * fix: reject instant LastRow across residual filters Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * test: expect after-merge selector in instant vector guards Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * docs: explain instant LastRow filter eligibility Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * fix: restrict instant LastRow to safe selector nodes Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * refactor: show LastRow merge mode directly in diagnostics Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> * test: refresh LastRow display in explain expectations Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> --------- Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com> |
||
 shuiyisong
|
b86da3d35f |
feat: support raw OTLP delta metrics (#8970)
* feat: support raw OTLP delta metrics Signed-off-by: shuiyisong <xixing.sys@gmail.com> * fix: fmt Signed-off-by: shuiyisong <xixing.sys@gmail.com> * test(promql): update sqlness results for normalized label matching Signed-off-by: shuiyisong <xixing.sys@gmail.com> * fix: derive temporality label from default column prefix Signed-off-by: shuiyisong <xixing.sys@gmail.com> * test(promql): add analyze coverage for delta temporality Signed-off-by: shuiyisong <xixing.sys@gmail.com> * fix(promql): scope label alignment to temporality marker Signed-off-by: shuiyisong <xixing.sys@gmail.com> * fix: handle count-only histograms and vector broadcasts Signed-off-by: shuiyisong <xixing.sys@gmail.com> * fix: exclude temporality marker from entity descriptions Signed-off-by: shuiyisong <xixing.sys@gmail.com> * fix: use a fixed label for OTLP aggregation temporality Signed-off-by: shuiyisong <xixing.sys@gmail.com> * fix(promql): preserve mixed-range semantics for raw delta Signed-off-by: shuiyisong <xixing.sys@gmail.com> --------- Signed-off-by: shuiyisong <xixing.sys@gmail.com> |
||
|
discord9
|
3493d2d0fb |
perf(servers)!: speed up Prometheus JSON response building with ryu and per-series entry reuse (#8815)
* fix(perf): align direct-SST CREATE TABLE with baked index metadata
The offline fixture generator (query_perf_fixture::direct_sst::
build_region_metadata) bakes greptime:inverted_index /
greptime:skipping_index field metadata into the region manifest for
tag/field columns, but create_table_sql emitted a bare CREATE TABLE
without those declarations. MergeScan's remote-schema validation then
failed on any tag/field projection (HTTP 500 'advertised remote stream
schema field mismatch'), breaking direct_readable_sst perf cases.
CREATE TABLE now declares the matching SKIPPING INDEX WITH
(granularity='1') / INVERTED INDEX column options. A round-trip test
proves the emitted SQL is parser-valid and yields the exact catalog
metadata.
Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com>
* perf(servers): speed up Prometheus JSON response building with ryu and per-series entry reuse
PrometheusJsonResponse::record_batches_to_data spends ~47% of its CPU
in f64::to_string() per sample and ~32% in IndexMap::entry() per row
(60s profile of concurrent query_range workloads, ~800k series).
- Replace f64::to_string() with ryu::Buffer::format_finite for finite
values (shortest round-trip, 3-5x faster); NaN/+Inf/-Inf keep the
previous std formatting so wire output is unchanged.
- Remember the previous row's label vector and entry index; query output
is clustered by series, so consecutive rows reuse the same IndexMap
entry via get_index_mut instead of rebuilding and hashing the label
vector (worst case adds one Vec comparison per series transition).
Also adds a query-regression case (prom_json_response) that measures the
real Prometheus HTTP range API path (/v1/prometheus/api/v1/query_range),
which is the only frontend path that builds the Prometheus JSON response
(TQL ANALYZE formats the SQL JSON shape instead), plus a prom_http query
kind in the regression runner.
Perf (aligned base
|
||
|
discord9
|
a7590f8174 |
perf(promql): avoid repeated scans in sliding range evaluation (#8646)
* perf(promql): use two pointers for sliding range boundaries Replace the stale cursor heuristic in RangeManipulateStream::calculate_range with monotonic left/right cursors. The old path rescanned each evaluation window (O(E x samples-per-window)) and could lose valid samples after sparse gaps or trailing empty windows. The two pointers keep strict monotonic progress, reducing boundary generation to O(N + E) while preserving (curr-range, curr] semantics, start/end shortening, and empty-window output. Controlled release benchmarks (fixed CPU, ABBA): - Public RangeManipulate wall time: ~28% faster at 1m/15s, ~66% at 5m/15s, ~96% at 1h/15s. - Warmed distributed TQL ANALYZE 1h queries: ~17-21% faster end to end; shorter windows stayed within run-order noise. Signed-off-by: discord9 <discord9@163.com> * perf(promql): specialize changes/resets with adaptive edge counting The generic range_fn macro slices, downcasts, and rescans every overlapping window for changes() and resets(). Replace the macro path for these two functions with hand-written UDF wrappers backed by a shared private edge-count kernel: direct raw-offset scans when requested edges are few, otherwise one global u64 edge prefix so each window is answered by a prefix difference. Behavior is preserved bit-for-bit, including raw null-buffer values, NaN semantics, signed zero, infinities, empty/singleton windows, independent timestamp/value offsets, arbitrary window layouts, and exact DataFusion error messages. The shared proc macro, planner, serializer, and other range functions are untouched. Controlled release benchmarks (fixed CPU, ABBA): - Dense sliding windows (k=4/20/240): 91.7-95.6% less public UDF wall time. - Low-coverage fallback (N=4096, 8 windows): 73.9-74.4% faster. - Warmed distributed TQL ANALYZE 5m/1h changes/resets: 12.1-19.7% client and 12.0-20.9% server latency improvement; controls stayed within drift. Signed-off-by: discord9 <discord9@163.com> * ci(query-regression): include PromQL range boundary case in defaults An audit of historical query-regression runs found zero range-query coverage: all 208 PromQL ANALYZE samples were bare selectors, so range evaluation could regress without CI noticing. Wire the promql_range_boundary case (introduced in #8646) into DEFAULT_CASES so label-triggered runs measure the range path. The case is cheap: a ~0.3s synthetic fixture and about a minute of query execution per base/candidate pass. Signed-off-by: discord9 <discord9@163.com> * chore(promql): address sliding range review nits Move test-only imports into their test modules and remove the unused pre-specialization changes and resets helpers. Signed-off-by: discord9 <discord9@163.com> * style(promql): apply pinned rustfmt Signed-off-by: discord9 <discord9@163.com> * test(promql): cover sparse range results Share the changes and resets test scaffolding while keeping their behavior oracles independent. Add an end-to-end sqlness regression for sparse samples, empty intermediate windows, and a valid trailing sample. Signed-off-by: discord9 <discord9@163.com> --------- Signed-off-by: discord9 <discord9@163.com> |
||
|
shuiyisong
|
f5f5d468bb |
ci: add OTLP trace ingestion regression testing (#8631)
* chore: update CI config Signed-off-by: shuiyisong <xixing.sys@gmail.com> * chore: add CI Signed-off-by: shuiyisong <xixing.sys@gmail.com> * chore: update CI config Signed-off-by: shuiyisong <xixing.sys@gmail.com> * ci: report otelgen runner diagnostics Signed-off-by: shuiyisong <xixing.sys@gmail.com> * chore: add script to draw result diagram Signed-off-by: shuiyisong <xixing.sys@gmail.com> --------- Signed-off-by: shuiyisong <xixing.sys@gmail.com> |
||
|
discord9
|
56addd0623 |
fix: stream remote analyze metrics while pending (#8405)
* fix: stream remote analyze metrics while pending Signed-off-by: discord9 <discord9@163.com> * test: verify flight metrics preserve pending batch Signed-off-by: discord9 <discord9@163.com> * fix: preserve direct SST perf queries in plans Signed-off-by: discord9 <discord9@163.com> * fix: bind flight metrics capability to query Signed-off-by: discord9 <discord9@163.com> --------- Signed-off-by: discord9 <discord9@163.com> |
||
|
discord9
|
02283b6ba0 |
test(perf): add remote write storage inspection (#8444)
* feat: add remote write value distributions Signed-off-by: discord9 <discord9@163.com> * test: extend remote write value distributions Signed-off-by: discord9 <discord9@163.com> * test: inspect remote write parquet storage Signed-off-by: discord9 <discord9@163.com> * test: normalize remote write perf fixtures Signed-off-by: discord9 <discord9@163.com> * test: add heavy remote write perf case Signed-off-by: discord9 <discord9@163.com> * test: use head greptime for read bench Signed-off-by: discord9 <discord9@163.com> * test: keep heavy remote write case local Signed-off-by: discord9 <discord9@163.com> * test: tune remote write perf smoke case Signed-off-by: discord9 <discord9@163.com> * test: fix query fixture import style Signed-off-by: discord9 <discord9@163.com> * test: cover remote write value distributions Signed-off-by: discord9 <discord9@163.com> * test: add integer counter perf case Signed-off-by: discord9 <discord9@163.com> --------- Signed-off-by: discord9 <discord9@163.com> |
||
|
discord9
|
12f83828b1 |
feat: add Prom remote-write query regression scenario (#8413)
* feat: add Prom remote-write query regression scenario Signed-off-by: discord9 <discord9@163.com> * test: add high-cardinality remote-write query case Signed-off-by: discord9 <discord9@163.com> * feat: chunk remote-write query regression loads Signed-off-by: discord9 <discord9@163.com> * test: use multi-day remote-write regression case Signed-off-by: discord9 <discord9@163.com> * fix: address query regression review comments Signed-off-by: discord9 <discord9@163.com> * ci: allow large query regression comments Signed-off-by: discord9 <discord9@163.com> --------- Signed-off-by: discord9 <discord9@163.com> |
||
|
discord9
|
c44f8da646 |
feat: add query regression perf harness (#8406)
* feat: add query regression perf harness Signed-off-by: discord9 <discord9@163.com> * feat: extend query regression cases Signed-off-by: discord9 <discord9@163.com> * ci: harden query regression workflows Signed-off-by: discord9 <discord9@163.com> * fix: address query regression review comments Signed-off-by: discord9 <discord9@163.com> * ci: limit query regression PR triggers Signed-off-by: discord9 <discord9@163.com> * ci: run full query regression case set Signed-off-by: discord9 <discord9@163.com> * refactor: model query regression scenarios Signed-off-by: discord9 <discord9@163.com> * fix: avoid unenforced query regression thresholds Signed-off-by: discord9 <discord9@163.com> --------- Signed-off-by: discord9 <discord9@163.com> |