Optimization in Numeric Terms Aggregation query for Large Bucket Counts

[vinaykpud]

Description

In bucket aggregations, data node sends topN bucket requested to the coordinator. The contract here is to return the buckets sorted by key but topN on the basis of value.
If the number of requested top-N buckets exceeds or close to the maximum bucket ordinal, making the use of a PriorityQueue for top-N selection inefficient or redundant. So we made following modifications:

1. use quickselect for topN if the requested size is greater than the 20% of the total buckets.
2. If the requested size is greater than the bucket size then return all the bucket.

Benchmarking test results :

#18703 (comment)

Related Issues

Resolves #18703
Related #18650

Check List

• Functionality includes testing.
• API changes companion pull request created, if applicable.
• Public documentation issue/PR created, if applicable.

By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.
For more information on following Developer Certificate of Origin and signing off your commits, please check here.

[github-actions]

❌ Gradle check result for 74295ec: FAILURE

Please examine the workflow log, locate, and copy-paste the failure(s) below, then iterate to green. Is the failure a flaky test unrelated to your change?

[github-actions]

❌ Gradle check result for 9f7c12d: FAILURE

Please examine the workflow log, locate, and copy-paste the failure(s) below, then iterate to green. Is the failure a flaky test unrelated to your change?

[github-actions]

❕ Gradle check result for e124eb1: UNSTABLE

Please review all flaky tests that succeeded after retry and create an issue if one does not already exist to track the flaky failure.

[codecov]

Codecov Report

❌ Patch coverage is 89.47368% with 10 lines in your changes missing coverage. Please review.
✅ Project coverage is 72.90%. Comparing base (c01ff89) to head (68e77e1).
⚠️ Report is 9 commits behind head on main.

Files with missing lines	Patch %	Lines
...egations/bucket/terms/BucketSelectionStrategy.java	92.95%	2 Missing and 3 partials ⚠️
...regations/bucket/terms/NumericTermsAggregator.java	76.47%	4 Missing ⚠️
...va/org/opensearch/search/DefaultSearchContext.java	80.00%	1 Missing ⚠️

Additional details and impacted files

@@             Coverage Diff              @@
##               main   #18702      +/-   ##
============================================
+ Coverage     72.89%   72.90%   +0.01%     
- Complexity    69318    69339      +21     
============================================
  Files          5642     5643       +1     
  Lines        318636   318712      +76     
  Branches      46107    46112       +5     
============================================
+ Hits         232254   232348      +94     
- Misses        67540    67569      +29     
+ Partials      18842    18795      -47     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[github-actions]

❌ Gradle check result for 66ccdaa: FAILURE

Please examine the workflow log, locate, and copy-paste the failure(s) below, then iterate to green. Is the failure a flaky test unrelated to your change?

[github-actions]

❌ Gradle check result for ced0d1b: FAILURE

Please examine the workflow log, locate, and copy-paste the failure(s) below, then iterate to green. Is the failure a flaky test unrelated to your change?

[github-actions]

❌ Gradle check result for 13c663b: FAILURE

Please examine the workflow log, locate, and copy-paste the failure(s) below, then iterate to green. Is the failure a flaky test unrelated to your change?

[github-actions]

✅ Gradle check result for adda83c: SUCCESS

[github-actions]

❌ Gradle check result for 0d3ecb4: FAILURE

Please examine the workflow log, locate, and copy-paste the failure(s) below, then iterate to green. Is the failure a flaky test unrelated to your change?

[github-actions]

❌ Gradle check result for 0d3ecb4: FAILURE

Please examine the workflow log, locate, and copy-paste the failure(s) below, then iterate to green. Is the failure a flaky test unrelated to your change?

[jainankitk]

Approach is similar to terms aggregation. Can you answer the questions in #18732 (review) ?

[github-actions]

✅ Gradle check result for 68e77e1: SUCCESS

[vinaykpud]

@jainankitk

Approach is similar to terms aggregation. Can you answer the questions in #18732 (review) ?

Sure,

1. I am curious how we arrived at 20% as the right threshold for choosing between pq approach vs quickselect?

Compared with different threshold starting from 10% and analyzed for which value QuickSelects performs better. I have added the results here:
#18702 (comment)

2. Does this have any memory usage implications when the size is above 20% of value count?

There is no much difference in the memory usage is observed for the size above 20%. When we use quickSelect we create an array with size equal to the bucketOrdinal and copy all buckets to it to perform the topN selection. Bellow link has JVM metrics comparison.

#18703 (comment)