[
https://issues.apache.org/jira/browse/HIVE-29733?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Attila Turoczy updated HIVE-29733:
----------------------------------
Labels: cloud (was: )
> Establish Upstream Hive Performance Testing
> -------------------------------------------
>
> Key: HIVE-29733
> URL: https://issues.apache.org/jira/browse/HIVE-29733
> Project: Hive
> Issue Type: New Feature
> Reporter: Attila Turoczy
> Priority: Major
> Labels: cloud
>
> Apache Hive currently does not have a consistently executed and publicly
> documented upstream performance test framework that can be used to identify
> query performance regressions, compare storage formats, and validate
> improvements across releases.
> We should establish a repeatable upstream performance-testing process based
> on TPC-DS queries, executed against both traditional *external tables* and
> {*}Apache Iceberg tables{*}.
> Define:
> * The benchmark datasets
> * Table formats and layouts
> * Cluster configuration
> * Execution engine configuration
> * Test frequency
> * Cost-control strategy
> * Result storage and comparison
> * Regression thresholds
> * The one-billion-row challenge
> * Client and test-runner environments
> The primary objective is to create a stable performance baseline that can be
> used by the Apache Hive community to detect regressions and validate
> performance improvements.
>
>
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
