Create and traverse avro schemas once per task (#296) #834
+164
−41
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
* Cleanup and isolate the schema perf changes * add an _
smcnamara2-stripe
changed the title
pengyu-hou
approved these changes
Sep 26, 2024
hzding621
approved these changes
Nov 19, 2024
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Outdated suggestions cannot be applied.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.
Summary
This adds a SchemaTraverser to the chronon row creation logic, which allows callers of Row.to to specify how the output schema should be constructed.
At the core, this code is centered around removing the
AvroConversions.fromChrononSchemacall in the lambda passed fromfromChrononRow.Why / Goal
The proximate goal is to allow faster Avro Schema construction, avoiding repeated calls to fromChrononSchema in a tight loop. This has resulted in a 90+% reduction in CPU usage during the reduce phase of the GroupByUpload aggregation.
At stripe, several of our largest GroupByUpload apps are now using 50-80% fewer vcore-seconds per run.
Test Plan
Checklist
Reviewers