move Batch deserialization out of the lock #17050
Merged
Conversation
This file contains 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
jsvd
commented
Feb 11, 2025
yaauie
reviewed
Feb 14, 2025
There was a problem hiding this comment.
Mechanically, this makes sense, and I see no problem with it. I can't think of any way that deferring the deserialization until outside of the lock would cause problems.
I would prefer for the intermediate object to be an internal detail of the Queue class, and to have a name that more-accurately describes what it is.
I have opened jsvd#9 to show what I mean by that.
jsvd
changed the title
Currently the deserialization is behind the readBatch's lock, so any large batch will take time deserializing, causing any other Queue writer (e.g. netty executor threads) and any other Queue reader (pipeline worker) to block. This commit moves the deserialization out of the lock, allowing multiple pipeline workers to deserialize batches concurrently. - add intermediate batch-holder from `Queue` methods - make the intermediate batch-holder a private inner class of `Queue` with a descriptive name `SerializedBatchHolder` Co-authored-by: Ry Biesemeyer <[email protected]>
jsvd
force-pushed
the
unlock_queue_deserialization
branch
from
b4a6109 to
a2e7667
Compare
|
💚 Build Succeeded
History
|
|
Thanks @mashhurs for the review and Joao for the work :) |
|
@logstashmachine backport 8.18 |
|
@logstashmachine backport 8.17 |
|
@logstashmachine backport 8.16 |
github-actions bot
pushed a commit
that referenced
this pull request
Feb 17, 2025
Currently the deserialization is behind the readBatch's lock, so any large batch will take time deserializing, causing any other Queue writer (e.g. netty executor threads) and any other Queue reader (pipeline worker) to block. This commit moves the deserialization out of the lock, allowing multiple pipeline workers to deserialize batches concurrently. - add intermediate batch-holder from `Queue` methods - make the intermediate batch-holder a private inner class of `Queue` with a descriptive name `SerializedBatchHolder` Co-authored-by: Ry Biesemeyer <[email protected]> (cherry picked from commit 637f447)
github-actions bot
pushed a commit
that referenced
this pull request
Feb 17, 2025
Currently the deserialization is behind the readBatch's lock, so any large batch will take time deserializing, causing any other Queue writer (e.g. netty executor threads) and any other Queue reader (pipeline worker) to block. This commit moves the deserialization out of the lock, allowing multiple pipeline workers to deserialize batches concurrently. - add intermediate batch-holder from `Queue` methods - make the intermediate batch-holder a private inner class of `Queue` with a descriptive name `SerializedBatchHolder` Co-authored-by: Ry Biesemeyer <[email protected]> (cherry picked from commit 637f447)
github-actions bot
pushed a commit
that referenced
this pull request
Feb 17, 2025
Currently the deserialization is behind the readBatch's lock, so any large batch will take time deserializing, causing any other Queue writer (e.g. netty executor threads) and any other Queue reader (pipeline worker) to block. This commit moves the deserialization out of the lock, allowing multiple pipeline workers to deserialize batches concurrently. - add intermediate batch-holder from `Queue` methods - make the intermediate batch-holder a private inner class of `Queue` with a descriptive name `SerializedBatchHolder` Co-authored-by: Ry Biesemeyer <[email protected]> (cherry picked from commit 637f447)
|
@logstashmachine backport 9.0 |
github-actions bot
pushed a commit
that referenced
this pull request
Feb 17, 2025
Currently the deserialization is behind the readBatch's lock, so any large batch will take time deserializing, causing any other Queue writer (e.g. netty executor threads) and any other Queue reader (pipeline worker) to block. This commit moves the deserialization out of the lock, allowing multiple pipeline workers to deserialize batches concurrently. - add intermediate batch-holder from `Queue` methods - make the intermediate batch-holder a private inner class of `Queue` with a descriptive name `SerializedBatchHolder` Co-authored-by: Ry Biesemeyer <[email protected]> (cherry picked from commit 637f447)
jsvd
added a commit
that referenced
this pull request
Feb 17, 2025
Currently the deserialization is behind the readBatch's lock, so any large batch will take time deserializing, causing any other Queue writer (e.g. netty executor threads) and any other Queue reader (pipeline worker) to block. This commit moves the deserialization out of the lock, allowing multiple pipeline workers to deserialize batches concurrently. - add intermediate batch-holder from `Queue` methods - make the intermediate batch-holder a private inner class of `Queue` with a descriptive name `SerializedBatchHolder` Co-authored-by: Ry Biesemeyer <[email protected]> (cherry picked from commit 637f447) Co-authored-by: João Duarte <[email protected]>
jsvd
added a commit
that referenced
this pull request
Feb 17, 2025
Currently the deserialization is behind the readBatch's lock, so any large batch will take time deserializing, causing any other Queue writer (e.g. netty executor threads) and any other Queue reader (pipeline worker) to block. This commit moves the deserialization out of the lock, allowing multiple pipeline workers to deserialize batches concurrently. - add intermediate batch-holder from `Queue` methods - make the intermediate batch-holder a private inner class of `Queue` with a descriptive name `SerializedBatchHolder` Co-authored-by: Ry Biesemeyer <[email protected]> (cherry picked from commit 637f447) Co-authored-by: João Duarte <[email protected]>
jsvd
added a commit
that referenced
this pull request
Feb 17, 2025
Currently the deserialization is behind the readBatch's lock, so any large batch will take time deserializing, causing any other Queue writer (e.g. netty executor threads) and any other Queue reader (pipeline worker) to block. This commit moves the deserialization out of the lock, allowing multiple pipeline workers to deserialize batches concurrently. - add intermediate batch-holder from `Queue` methods - make the intermediate batch-holder a private inner class of `Queue` with a descriptive name `SerializedBatchHolder` Co-authored-by: Ry Biesemeyer <[email protected]> (cherry picked from commit 637f447) Co-authored-by: João Duarte <[email protected]>
jsvd
added a commit
that referenced
this pull request
Feb 17, 2025
Currently the deserialization is behind the readBatch's lock, so any large batch will take time deserializing, causing any other Queue writer (e.g. netty executor threads) and any other Queue reader (pipeline worker) to block. This commit moves the deserialization out of the lock, allowing multiple pipeline workers to deserialize batches concurrently. - add intermediate batch-holder from `Queue` methods - make the intermediate batch-holder a private inner class of `Queue` with a descriptive name `SerializedBatchHolder` Co-authored-by: Ry Biesemeyer <[email protected]> (cherry picked from commit 637f447) Co-authored-by: João Duarte <[email protected]>
|
@logstashmachine backport 8.x |
github-actions bot
pushed a commit
that referenced
this pull request
Feb 18, 2025
Currently the deserialization is behind the readBatch's lock, so any large batch will take time deserializing, causing any other Queue writer (e.g. netty executor threads) and any other Queue reader (pipeline worker) to block. This commit moves the deserialization out of the lock, allowing multiple pipeline workers to deserialize batches concurrently. - add intermediate batch-holder from `Queue` methods - make the intermediate batch-holder a private inner class of `Queue` with a descriptive name `SerializedBatchHolder` Co-authored-by: Ry Biesemeyer <[email protected]> (cherry picked from commit 637f447)
jsvd
added a commit
that referenced
this pull request
Feb 18, 2025
Currently the deserialization is behind the readBatch's lock, so any large batch will take time deserializing, causing any other Queue writer (e.g. netty executor threads) and any other Queue reader (pipeline worker) to block. This commit moves the deserialization out of the lock, allowing multiple pipeline workers to deserialize batches concurrently. - add intermediate batch-holder from `Queue` methods - make the intermediate batch-holder a private inner class of `Queue` with a descriptive name `SerializedBatchHolder` Co-authored-by: Ry Biesemeyer <[email protected]> (cherry picked from commit 637f447) Co-authored-by: João Duarte <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
Release Notes
Discussion
currently the deserialization is behind the readBatch's lock, so any large batch will take time deserializing, causing any other Queue writer (e.g. netty executor threads) and any other Queue reader (pipeline worker) to block.
This commit moves the deserialization out of the lock, allowing multiple pipeline workers to deserialize batches concurrently.
To test, edit the generator input to produce 100kb messages, and run:
bin/logstash -e "input { generator { threads => 5 } } output { null {} }"with PQ enabled.This PR should shows about a 4-5x performance improvement for 10 worker pipeline: