Alternative producer implementation #1285
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
Refactoring of the producer so that it handles errors per record.
domdorn
reviewed
Jul 17, 2024
| (metadata: RecordMetadata, err: Exception) => | ||
| unsafeRun { | ||
| if (err == null) done.succeed(metadata) | ||
| else done.fail(err) |
There was a problem hiding this comment.
if we change this to done.fail((record,err)) we could implement a retry strategy in 494
There was a problem hiding this comment.
I like the idea. But are you sure you want to do the retries record by record?
There was a problem hiding this comment.
I think that is the only way to not send messages again that have been previously accepted, e.g. prevent double delivery?
There was a problem hiding this comment.
The other option is to do something after line 489. There you can records.zip(results) and filter those that failed. The failed records can then be retried together. The challenge is mostly to write elegant code that merges the original results with the retry results.
svroonland
approved these changes
Jul 18, 2024
svroonland
pushed a commit
that referenced
this pull request
Aug 10, 2024
Refactoring of the producer so that it handles errors per record.
ytalashko
added a commit
to ytalashko/zio-kafka
that referenced
this pull request
Aug 21, 2024
This reverts commit fd40816.
erikvanoosten
pushed a commit
that referenced
this pull request
Aug 21, 2024
During running services with the new version of library 2.8.1, I noticed huge increase in messages production time to kafka. Some quick rough tests shoving me around 40x-100x times increase in amount of time taken to `produceChunk` on even 1-10 records chunks comparing to version 2.8.0 (or also to version 2.8.1 with the changes in this MR). Please, note, it is not a proper benchmarks. Also, to note, Im using Scala version 3.3.1. This MR just reverts two MRs updating `ProducersendFromQueue` implementation: - #1272 - #1285 Not sure if it's possible to revert two MRs at a time (with a single one for revert), so created this one. I haven't researched yet which exact change/changes are causing such performance degradation. I would suggest the next steps: - confirm the problem exists - reverting to the previous implementation (the one from 2.8.0/this MR) - release fixed version (to allow users have a nicely working version) - investigate & fix problem from the #1272 and/or #1285 It is only suggestions on the approach, feel free to ignore them. Also, feel free to modify/ignore this MR and treat it as an issue.
erikvanoosten
added a commit
that referenced
this pull request
Aug 24, 2024
Refactoring of the producer so that it handles errors per record. These changes were part of 2.8.1 (via #1285 and #1272) but were reverted in 2.8.2 (via #1304) after reports of performance regression. This is a retry to make it easier to run the benchmarks and make further improvements. This reverts commit 3393fbf.
erikvanoosten
added a commit
that referenced
this pull request
Aug 24, 2024
Refactoring of the producer so that it handles errors per record. These changes were part of 2.8.1 (via #1285 and #1272) but were reverted in 2.8.2 (via #1304) after reports of performance regression. This is a retry to make it easier to run the benchmarks and make further improvements. This reverts commit 3393fbf.
erikvanoosten
added a commit
that referenced
this pull request
Aug 25, 2024
Refactoring of the producer so that it handles errors per record. These changes were part of 2.8.1 (via #1285 and #1272) but were reverted in 2.8.2 (via #1304) after reports of performance regression. This is a retry to make it easier to run the benchmarks and make further improvements. This reverts commit 3393fbf.
erikvanoosten
added a commit
that referenced
this pull request
Aug 26, 2024
Refactoring of the producer so that it handles errors per record. These changes were part of 2.8.1 (via #1285 and #1272) but were reverted in 2.8.2 (via #1304) after reports of performance regression. This is a retry to make it easier to run the benchmarks and make further improvements. This reverts commit 3393fbf.
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.
Refactoring of the producer so that it handles errors per record.