Decrease size of _tsids created by TsidBuilder#133631
Merged
felixbarny merged 4 commits intoelastic:mainfrom Sep 15, 2025
Merged
Decrease size of _tsids created by TsidBuilder#133631felixbarny merged 4 commits intoelastic:mainfrom
felixbarny merged 4 commits intoelastic:mainfrom
Conversation
Based on testing, this has marginal impact on storage but decreases the size of the _tsids from 21B-36B to 18B-21B.
elasticsearchmachine
added
needs:triage
elasticsearchmachine
added
Team:StorageEngine
and removed
needs:triage
Collaborator
|
Pinging @elastic/es-storage-engine (Team:StorageEngine) |
kkrik-es
reviewed
Aug 27, 2025
server/src/test/java/org/elasticsearch/cluster/routing/TsidBuilderTests.java
Outdated
Show resolved
Hide resolved
kkrik-es
approved these changes
Aug 27, 2025
kkrik-es
reviewed
Aug 27, 2025
| * <ul> | ||
| * <li> | ||
| * A hash of the dimension field names (4 bytes). | ||
| * A hash of the dimension field names (1 byte). |
Contributor
There was a problem hiding this comment.
What happens if we skip this? is it worth experimenting with? It'd be nice to have a tsid that's always the same size e.g. 20b.
Member
Author
There was a problem hiding this comment.
Storage is a little worse, but not much. We may be overfitting to a specific dataset, though. We could also have a fixed 4 bytes instead of a variable 1-4 bytes for the values similarity hash.
This was referenced Aug 28, 2025
Member
Author
|
The impact of this on the
I don't understand why the throughput has regressed compared to #132566 (comment). It should be similar. I suspect that the throughput is quite volatile and can't easily be compared run-to-run. But before going with this approach, we should double-check this. The storage size should be stable as the track includes a force-merge step. Full Rally comparison ...Esbench params ... |
Contributor
|
Do we have results for 16B too? |
Member
Author
|
Yes, they're here: #133706 (comment) |
Contributor
|
Seems like a good tradeoff, @martijnvg wdyt? |
Member
Author
|
I'm trying to re-run the benchmarks with beefier instances that are closer to the nightly configuration. The variance for the throughput seems higher than expected. Having some issues with esbench atm. |
martijnvg
reviewed
Sep 11, 2025
Member
Author
|
Merging after the latest round of testing: #132566 |
mridula-s109
pushed a commit
to mridula-s109/elasticsearch
that referenced
this pull request
Sep 17, 2025
Based on testing, this has marginal impact on storage but decreases the size of the _tsids from 21B-36B to 18B-21B.
gmjehovich
pushed a commit
to gmjehovich/elasticsearch
that referenced
this pull request
Sep 18, 2025
Based on testing, this has marginal impact on storage but decreases the size of the _tsids from 21B-36B to 18B-21B.
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
4 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.
Based on testing, this has marginal impact on storage but decreases the size of the _tsids from 21B-36B to 18B-21B.
Follow-up from #133344
Part of #132566