feat(tsdb): allow using ST capable XOR chunks - retain format on read by krajorama · Pull Request #18013 · prometheus/prometheus

krajorama · 2026-02-04T23:02:42Z

Two big changes

Use the new XOROPTST chunk format optionally for appending samples
When compacting/querying chunks via the populateWithDelChunkSeriesIterator take into account if the source samples have ST==0 or not.

The second change is there to make sure we don't convert existing chunks to a new format accidentally.
It does mean that we're converting to the old format some chunks that need deletion.

Based on #18103
Finishes #17791 for floats.

Which issue(s) does the PR fix:

Does this PR introduce a user-facing change?

NONE

krajorama · 2026-02-05T05:58:33Z

/prombench main

prombot · 2026-02-05T05:58:36Z

⏱️ Welcome to Prometheus Benchmarking Tool. ⏱️

Compared versions: PR-18013 and main

After the successful deployment (check status here), the benchmarking results can be viewed at:

Available Commands:

To restart benchmark: /prombench restart main
To stop benchmark: /prombench cancel
To print help: /prombench help

krajorama · 2026-02-05T12:41:50Z

/prombench cancel

prombot · 2026-02-05T12:41:53Z

Benchmark cancel is in progress.

krajorama · 2026-02-05T12:57:46Z

/prombench main

prombot · 2026-02-05T12:57:50Z

⏱️ Welcome to Prometheus Benchmarking Tool. ⏱️

Compared versions: PR-18013 and main

After the successful deployment (check status here), the benchmarking results can be viewed at:

Available Commands:

To restart benchmark: /prombench restart main
To stop benchmark: /prombench cancel
To print help: /prombench help

krajorama · 2026-02-05T13:26:37Z


 	for currentValueType := firstValueType; currentValueType != chunkenc.ValNone; currentValueType = p.currDelIter.Next() {
+		var (
+			newChunk chunkenc.Chunk


Note to reviewers: I think this was a bug, since we never cleared newChunk . However it only hits if a series went from histogram to float and the histogram needed a new chunk right before this.

bwplotka · 2026-02-06T10:03:42Z

I like what I see (:

krajorama · 2026-02-06T18:30:30Z

/prombench cancel

prombot · 2026-02-06T18:30:34Z

Benchmark cancel is in progress.

bwplotka

First glance, one question around retaining implementaiton

bwplotka · 2026-02-18T09:34:25Z

 		// For the first sample, the following condition will always be true as
 		// ValNone != ValFloat | ValHistogram | ValFloatHistogram.
-		if currentValueType != prevValueType {
+		encoding := p.currDelIter.Encoding()


Sounds like this is the only place where we actually use chunkenc.Iterator Encoding.

Is there another way we should decide what new chunk to use? One is your previous solution - based on flag, with different semantics.

Here I'm worried about interface explosion and leaking abstraction. In majority of chunk iterator implementations returning Encoding() makes no sense as it's not a chunk even. Avoiding this would be amazing.

So flag change is one way (and not retaking).

But it feels this currDelIter will always have same a chunk as a source - so why not persisting what was the original encoding (ST with or without, or even full encoding) in

prometheus/tsdb/querier.go

Line 647 in 8d491cc

p.currMeta = p.metas[p.i]

Side question is around this logic for value types. How a single chunk can have a mixed value types? When this can happen? I think currDelIter is only formed from a single chunk, looking on func (p *populateWithDelChunkSeriesIterator) Next() bool { and func (p *populateWithDelGenericSeriesIterator) next(copyHeadChunk bool) bool

I've updated the code to not use this new interface, rather take into account ST==0 or not.
This does mean that some chunks with ST support that we return on query or write at compaction, might get turned into the original non ST capable chunks if the written ST is actually 0 AND the chunk need recoding due to:

vertical merge is needed due to overlapping in-order and out-of-order chunks

See PR description for uptodate info on details. Signed-off-by: György Krajcsovits <[email protected]>

When querying chunks from storage, take into account the source encoding instead of losing ST. Signed-off-by: György Krajcsovits <[email protected]>

Generated with Claude Opus 4.5. Signed-off-by: György Krajcsovits <[email protected]>

Also add tests by Claude Opus 4.5. Signed-off-by: György Krajcsovits <[email protected]>

Generated with Claude Opus 4.5. Signed-off-by: György Krajcsovits <[email protected]>

Signed-off-by: György Krajcsovits <[email protected]>

bwplotka · 2026-02-19T10:26:09Z

 }

-func (v ValueType) ChunkEncoding() Encoding {
+func (v ValueType) ChunkEncoding(storeST bool) Encoding {


much cleaner, thanks!

bwplotka

Reviewed in detail.

Generally logic and semantics LGTM, thanks for this! I'd missed OOO impl in my attempt, good findings. I love the new ST detection on querier, so much cleaner

Suggesting changes around the maintenance toil reducing opportunities

Let's not increase function signatures with new feature flag if we won't need to. E.g. chunkOptions can be reused. If we can't reuse opts, I propose we add ff as the last argument because this likely will be removed eventually and it's not a first thing you need to know when reading signatures.
Kill AppenderV1 test for STs

Some styling nits that I deeply care about and I know you disagree on some 🙃 . World will not burn if you ignore me there:

Odd // AppendV2 signature: (ref, labels, st, t, v, h, fh, opts) comments
Trailing dots missing on comments.
Table tests should use slices IMO (unless we change globally to maps)

bwplotka · 2026-02-19T10:44:47Z

 			}

-			if _, chunkCreated := ms.append(s.T, s.V, 0, appendChunkOpts); chunkCreated {
+			if _, chunkCreated := ms.append(wp.storeST, s.ST, s.T, s.V, 0, appendChunkOpts); chunkCreated {


ditto -it would be better if ff would be in appendChunkOpts (fewer changes, easier to remove later)

bwplotka · 2026-02-19T10:49:04Z

+	}
+
+loop:
+	for vt := valueType; vt != chunkenc.ValNone; vt = p.currDelIter.Next() {


this is now way cleaner!

bwplotka · 2026-02-19T10:50:45Z

 		// For the first sample, the following condition will always be true as
 		// ValNone != ValFloat | ValHistogram | ValFloatHistogram.
-		if currentValueType != prevValueType {
+		// Also if we need to store start time (ST), but the current chunk is


super clean!

Signed-off-by: György Krajcsovits <[email protected]>

bwplotka

I think you agree with my comments but they were not fully addressed (in all related places). I pointed to the places when this happens.

Otherwise LGTM

By AI Signed-off-by: György Krajcsovits <[email protected]>

Round 2 for cutNewOOOHeadChunk, mmapCUrrentOOOHeadChunk. Signed-off-by: György Krajcsovits <[email protected]>

…ncoding` (#18062) * feat(tsdb/chunkenc): add float chunk format with start timestamp support Signed-off-by: György Krajcsovits <[email protected]> * optimize code path and layout Signed-off-by: György Krajcsovits <[email protected]> * make new format usable in head Signed-off-by: György Krajcsovits <[email protected]> * fix issue with seeking to last sample again Signed-off-by: György Krajcsovits <[email protected]> * fix iterator benchmark for chunks not supporting ST Signed-off-by: György Krajcsovits <[email protected]> * reduce footprint of the xoroptst chunk iterator object It was 80 bytes with a lot of padding compared to the 56 bytes of the original xor chunk iterator. Made it 64 bytes, tightly packed. Signed-off-by: György Krajcsovits <[email protected]> * Fix benchmark expectations on ST in interator Signed-off-by: György Krajcsovits <[email protected]> * add inclusive delta test case Signed-off-by: György Krajcsovits <[email protected]> * make testcases independent of order Signed-off-by: György Krajcsovits <[email protected]> * drop unused code Signed-off-by: György Krajcsovits <[email protected]> * Drop commented out line Signed-off-by: György Krajcsovits <[email protected]> * documentation Signed-off-by: György Krajcsovits <[email protected]> * Small simplification in the doc Signed-off-by: György Krajcsovits <[email protected]> * Add delta st inclusive test case for random vt Signed-off-by: György Krajcsovits <[email protected]> * Switch to delta of difference of st to prev t from delta of delta of st. Signed-off-by: György Krajcsovits <[email protected]> * Write ST after T and V so we can write a single bit on the second sample Signed-off-by: György Krajcsovits <[email protected]> * verify chunk sample len function Signed-off-by: György Krajcsovits <[email protected]> * Reduce size of first st stored a little Signed-off-by: György Krajcsovits <[email protected]> * test the case where st equals the t Signed-off-by: György Krajcsovits <[email protected]> * add st equal t to bechmarks Signed-off-by: György Krajcsovits <[email protected]> * test(chunkenc): test that appender can contonue chunks Test that initializing a chunk appender from an existing chunk works correctly. Signed-off-by: György Krajcsovits <[email protected]> * fix(chunkenc): bug in initializing appender on existing chunk Signed-off-by: György Krajcsovits <[email protected]> * Add cases with jitter in the start time as well Signed-off-by: György Krajcsovits <[email protected]> * tsdb: ST-in-WAL: Counter implementation and benchmarks (#17671) Initial implementation of #17790. Only implements ST-per-sample for Counters. Tests and benchmarks updated. Note: This increases the size of the RefSample object for all users, whether st-per-sample is turned on or not. Signed-off-by: Owen Williams <[email protected]> * refactor: sed enableStStorage/enableSTStorage Signed-off-by: bwplotka <[email protected]> * feat[scrape]: add ST parsing support to scrape AppenderV2 flow (#18103) Signed-off-by: bwplotka <[email protected]> * feat(tsdb): change head opt EnableSTStorage to atomic (#18107) In downstream projects this needs to be set dynamically per tenant. Signed-off-by: György Krajcsovits <[email protected]> * Merge pull request #18108 from prometheus/bwplotka/fix scrape: add tests for ST appending; add warnings for ST feature flag users around _created drop * refact(tsdb): trivial rename (#18109) Signed-off-by: György Krajcsovits <[email protected]> * fix(tsdb): missing passing head option to wal/wbl write (#18113) Signed-off-by: György Krajcsovits <[email protected]> * feat(tsdb): allow using ST capable XOR chunks - retain format on read (#18013) * feat(tsdb): allow appending to ST capable XOR chunk optionally Only for float samples as of now. Supports for in-order and out-of-order samples. Make sure that on readout the ST capable chunks are returned automatically. When the chunks are returned as is, this is trivially true. When a chunk needs to be re-coded due to deletion (tombstone) markers, we take the encoding of the original chunk. When a chunk needs to be created from overlapping chunks, we observe whether ST is zero or not and create the new chunk based on that. Signed-off-by: György Krajcsovits <[email protected]> * fix test after merge Signed-off-by: bwplotka <[email protected]> * feat: RW2 sending ST support Signed-off-by: bwplotka <[email protected]> tmp Signed-off-by: bwplotka <[email protected]> * tests: test ST in a cheapest way possible Signed-off-by: bwplotka <[email protected]> * tests: add bench CLI recommended invokations Signed-off-by: bwplotka <[email protected]> * fixed tests after rebase Signed-off-by: bwplotka <[email protected]> * feat(chunkenc): replace xoroptst chunk encoding with xor2 XOR2 is based on #18238 With additional ST support. Signed-off-by: György Krajcsovits <[email protected]> * feat: add compliance RW sender test for agent Signed-off-by: bwplotka <[email protected]> * feat(agent): add support for appending ST Signed-off-by: bwplotka <[email protected]> * replace stray xoroptst words Signed-off-by: György Krajcsovits <[email protected]> * Apply suggestions from code review Co-authored-by: Copilot <[email protected]> Signed-off-by: Bartlomiej Plotka <[email protected]> * post merge conflict fixes Signed-off-by: bwplotka <[email protected]> * feat(tsdb): register st_storage in feature API Register the st-storage feature flag in the feature registry via the TSDB options, consistent with how other TSDB features like exemplar_storage and delayed_compaction are registered. Signed-off-by: György Krajcsovits <[email protected]> Coded with Claude Sonnet 4.6. * Document xor2-encoding feature flag Signed-off-by: Carrie Edwards <[email protected]> * Add xor2-encoding feature flag Signed-off-by: Carrie Edwards <[email protected]> * Update CHANGELOG Signed-off-by: Carrie Edwards <[email protected]> * Fix linting Signed-off-by: Carrie Edwards <[email protected]> * Remove setting of xor2 encoding option in db open Signed-off-by: Carrie Edwards <[email protected]> * Fix tests Signed-off-by: Carrie Edwards <[email protected]> * Fix linting Signed-off-by: Carrie Edwards <[email protected]> * Update feature flag description Signed-off-by: Carrie Edwards <[email protected]> * Update comments and feature flag description Signed-off-by: Carrie Edwards <[email protected]> * Update documentation for st-storage feature Signed-off-by: Carrie Edwards <[email protected]> * st: disconnect st-storage with xor2-encoding given planned experiments (#18316) * st: disconnect st-storage with xor2-encoding given planned experiments Signed-off-by: bwplotka <[email protected]> * Update docs/feature_flags.md Co-authored-by: George Krajcsovits <[email protected]> Signed-off-by: Bartlomiej Plotka <[email protected]> * Update docs/feature_flags.md Co-authored-by: George Krajcsovits <[email protected]> Signed-off-by: Bartlomiej Plotka <[email protected]> * Update docs/feature_flags.md Co-authored-by: George Krajcsovits <[email protected]> Signed-off-by: Bartlomiej Plotka <[email protected]> * Update docs/feature_flags.md Co-authored-by: George Krajcsovits <[email protected]> Signed-off-by: Bartlomiej Plotka <[email protected]> --------- Signed-off-by: bwplotka <[email protected]> Signed-off-by: Bartlomiej Plotka <[email protected]> Co-authored-by: George Krajcsovits <[email protected]> --------- Signed-off-by: György Krajcsovits <[email protected]> Signed-off-by: Ganesh Vernekar <[email protected]> Signed-off-by: Bryan Boreham <[email protected]> Signed-off-by: Aleksandr Smirnov <[email protected]> Signed-off-by: Mohammad Abbasi <[email protected]> Signed-off-by: matt-gp <[email protected]> Signed-off-by: Ian Kerins <[email protected]> Signed-off-by: SuperQ <[email protected]> Signed-off-by: dependabot[bot] <[email protected]> Signed-off-by: Arve Knudsen <[email protected]> Signed-off-by: ffgan <[email protected]> Signed-off-by: Patryk Prus <[email protected]> Signed-off-by: Owen Williams <[email protected]> Signed-off-by: bwplotka <[email protected]> Signed-off-by: 3Juhwan <[email protected]> Signed-off-by: Sammy Tran <[email protected]> Signed-off-by: Casie Chen <[email protected]> Signed-off-by: Dan Cech <[email protected]> Signed-off-by: kakabisht <[email protected]> Signed-off-by: Jeanette Tan <[email protected]> Signed-off-by: Divyansh Mishra <[email protected]> Signed-off-by: Varun Chawla <[email protected]> Signed-off-by: Martin Valiente Ainz <[email protected]> Signed-off-by: Kyle Eckhart <[email protected]> Signed-off-by: Julien Pivotto <[email protected]> Signed-off-by: Matthieu MOREL <[email protected]> Signed-off-by: Linas Medziunas <[email protected]> Signed-off-by: Björn Rabenstein <[email protected]> Signed-off-by: beorn7 <[email protected]> Signed-off-by: Sayuru <[email protected]> Signed-off-by: Matt <[email protected]> Signed-off-by: Bartlomiej Plotka <[email protected]> Signed-off-by: Carrie Edwards <[email protected]> Co-authored-by: Bartlomiej Plotka <[email protected]> Co-authored-by: Ganesh Vernekar <[email protected]> Co-authored-by: Bryan Boreham <[email protected]> Co-authored-by: Sasha <[email protected]> Co-authored-by: Mohammad Abbasi <[email protected]> Co-authored-by: matt-gp <[email protected]> Co-authored-by: Ian Kerins <[email protected]> Co-authored-by: SuperQ <[email protected]> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> Co-authored-by: Arve Knudsen <[email protected]> Co-authored-by: Julien <[email protected]> Co-authored-by: ffgan <[email protected]> Co-authored-by: Patryk Prus <[email protected]> Co-authored-by: Ganesh Vernekar <[email protected]> Co-authored-by: Joe Adams <[email protected]> Co-authored-by: renovate[bot] <29139614+renovate[bot]@users.noreply.github.com> Co-authored-by: Owen Williams <[email protected]> Co-authored-by: 3Juhwan <[email protected]> Co-authored-by: Casie Chen <[email protected]> Co-authored-by: Dan Cech <[email protected]> Co-authored-by: hridyesh bisht <[email protected]> Co-authored-by: zenador <[email protected]> Co-authored-by: Divyansh Mishra <[email protected]> Co-authored-by: Varun Chawla <[email protected]> Co-authored-by: Martin Valiente Ainz <[email protected]> Co-authored-by: Kyle Eckhart <[email protected]> Co-authored-by: Matthieu MOREL <[email protected]> Co-authored-by: Linas Medžiūnas <[email protected]> Co-authored-by: Björn Rabenstein <[email protected]> Co-authored-by: beorn7 <[email protected]> Co-authored-by: Sayuru <[email protected]> Co-authored-by: Copilot <[email protected]> Co-authored-by: Carrie Edwards <[email protected]>

krajorama requested review from a team, alexgreenbank, bwplotka, codesome, cstyan, jesusvazquez, roidelapluie and tomwilkie as code owners February 4, 2026 23:02

krajorama removed request for alexgreenbank and tomwilkie February 4, 2026 23:03

krajorama marked this pull request as draft February 4, 2026 23:04

prombot added the prombench label Feb 5, 2026

krajorama force-pushed the krajo/st-opt-append-to-chunk branch from 72f1857 to 06546bb Compare February 5, 2026 07:33

krajorama commented Feb 5, 2026

View reviewed changes

krajorama changed the title ~~feat(tsdb): allow using ST capable XOR chunk optionally~~ feat(tsdb): allow using ST capable XOR chunks - retain format on read Feb 5, 2026

Base automatically changed from krajo/st-in-xorchunk to feature/start-time February 6, 2026 18:28

bwplotka mentioned this pull request Feb 17, 2026

feat(tsdb): allow using ST capable XOR chunks - set format on read #18026

Closed

krajorama force-pushed the krajo/st-opt-append-to-chunk branch from 9cf544c to 4ac8ffb Compare February 17, 2026 18:15

krajorama changed the base branch from feature/start-time to bwplotka/scrape-parsest February 17, 2026 18:15

krajorama force-pushed the krajo/st-opt-append-to-chunk branch 2 times, most recently from b4b2f9d to e4c96d5 Compare February 17, 2026 19:15

krajorama marked this pull request as ready for review February 17, 2026 19:17

Base automatically changed from bwplotka/scrape-parsest to feature/start-time February 18, 2026 09:15

krajorama force-pushed the krajo/st-opt-append-to-chunk branch from e4c96d5 to cb90f9f Compare February 18, 2026 10:01

bwplotka reviewed Feb 18, 2026

View reviewed changes

bwplotka force-pushed the feature/start-time branch from 7f5534f to e25dc28 Compare February 18, 2026 11:20

krajorama added 11 commits February 18, 2026 18:50

feat(tsdb): allow appending to ST capable XOR chunk optionally

e40f988

See PR description for uptodate info on details. Signed-off-by: György Krajcsovits <[email protected]>

feat(tsdb): query of overlapping chunks with ST by source encoding

7c295db

When querying chunks from storage, take into account the source encoding instead of losing ST. Signed-off-by: György Krajcsovits <[email protected]>

test: ooo head insert tests for ST

5ada7a4

Generated with Claude Opus 4.5. Signed-off-by: György Krajcsovits <[email protected]>

fix(tsdb): extra chunk creation by using wrong default encoding

a5394a8

Also add tests by Claude Opus 4.5. Signed-off-by: György Krajcsovits <[email protected]>

test(tsdb): test ToEncodedChunks with ST

5592b80

Generated with Claude Opus 4.5. Signed-off-by: György Krajcsovits <[email protected]>

test(tsdb): tests for populateWithDelGenericSeriesIterator

d2a17e7

Signed-off-by: György Krajcsovits <[email protected]>

feat(tsdb): adopt head append changes from 18026

7991bcb

Signed-off-by: György Krajcsovits <[email protected]>

fix(chunkenc): get rid of Iterator.Encoding()

cf929d6

Signed-off-by: György Krajcsovits <[email protected]>

fix(test): removed wrong function.

8e2169f

Signed-off-by: György Krajcsovits <[email protected]>

fix(tsdb): wal/wbl do pass ST when requested

1f0cc81

Signed-off-by: György Krajcsovits <[email protected]>

fix(tsdb): remove populateCurrForSingleChunk complication

1ef2063

Signed-off-by: György Krajcsovits <[email protected]>

krajorama force-pushed the krajo/st-opt-append-to-chunk branch from 5bce7a8 to 1ef2063 Compare February 18, 2026 17:56

refact(tsdb): reuse variable instead of loading from head

9bebcdf

Signed-off-by: György Krajcsovits <[email protected]>

bwplotka reviewed Feb 19, 2026

View reviewed changes

bwplotka requested changes Feb 19, 2026

View reviewed changes

krajorama added 4 commits February 19, 2026 14:05

fix(tsdb): fix style and remove useless comments.

df07a49

Signed-off-by: György Krajcsovits <[email protected]>

fix(tsdb): remove TODO and revert order of conditions

175e103

Signed-off-by: György Krajcsovits <[email protected]>

refact(tsdb): use chunkOpts to pass storeST to functions

bc2d0bc

Signed-off-by: György Krajcsovits <[email protected]>

refact(tsdb): move storeST to last argument of ToEncodedChunks

b5f4d07

Signed-off-by: György Krajcsovits <[email protected]>

bwplotka approved these changes Feb 19, 2026

View reviewed changes

krajorama added 2 commits February 19, 2026 16:11

Fix comments

3d04477

By AI Signed-off-by: György Krajcsovits <[email protected]>

refact(tsdb): use chunkOpts for passing storeST and chunkDiskMapper

b6b702e

Round 2 for cutNewOOOHeadChunk, mmapCUrrentOOOHeadChunk. Signed-off-by: György Krajcsovits <[email protected]>

krajorama merged commit 223f016 into feature/start-time Feb 20, 2026
48 of 49 checks passed

krajorama deleted the krajo/st-opt-append-to-chunk branch February 20, 2026 08:15

Uh oh!

Conversation

krajorama commented Feb 4, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Which issue(s) does the PR fix:

Does this PR introduce a user-facing change?

Uh oh!

krajorama commented Feb 5, 2026

Uh oh!

prombot commented Feb 5, 2026

Uh oh!

krajorama commented Feb 5, 2026

Uh oh!

prombot commented Feb 5, 2026

Uh oh!

krajorama commented Feb 5, 2026

Uh oh!

prombot commented Feb 5, 2026

Uh oh!

krajorama Feb 5, 2026

Choose a reason for hiding this comment

Uh oh!

bwplotka commented Feb 6, 2026

Uh oh!

krajorama commented Feb 6, 2026

Uh oh!

prombot commented Feb 6, 2026

Uh oh!

bwplotka left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

bwplotka Feb 18, 2026

Choose a reason for hiding this comment

Uh oh!

krajorama Feb 18, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

bwplotka Feb 19, 2026

Choose a reason for hiding this comment

Uh oh!

bwplotka left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

bwplotka Feb 19, 2026

Choose a reason for hiding this comment

Uh oh!

krajorama Feb 19, 2026

Choose a reason for hiding this comment

Uh oh!

Uh oh!

bwplotka Feb 19, 2026

Choose a reason for hiding this comment

Uh oh!

bwplotka Feb 19, 2026

Choose a reason for hiding this comment

Uh oh!

bwplotka left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

krajorama commented Feb 4, 2026 •

edited

Loading

krajorama Feb 18, 2026 •

edited

Loading

bwplotka left a comment •

edited

Loading