Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[BUG] Date Histogram aggregation and it's variations have increased latency with OS 3.0/Lucene10 upgrade #17405

Open
expani opened this issue Feb 20, 2025 · 0 comments
Labels

Comments

@expani
Copy link
Contributor

expani commented Feb 20, 2025

Describe the bug

OS 3.0 * OS 2.19 indicates that the OpenSearch server was running 3.0 whereas the index used was created in OS 2.19. This was done to eliminate any suspicion of the bug arising from an indexing change in Lucene.

Latency and service time has increased irrespective of OS 3.0 or OS 2.19 index used with OS 3.0 server

Metric Name Operation/Query name OS 2.19 OS 3.0 OS 3.0 * OS 2.19 Unit
50th percentile latency date_histogram_minute_agg 40.6969 41.6565 43.0194 ms
90th percentile latency date_histogram_minute_agg 41.463 43.1025 43.9114 ms
99th percentile latency date_histogram_minute_agg 43.2176 45.2824 45.1605 ms
100th percentile latency date_histogram_minute_agg 43.4249 51.6785 51.1855 ms
50th percentile service time date_histogram_minute_agg 39.3274 40.1942 41.8148 ms
90th percentile service time date_histogram_minute_agg 39.9173 41.4967 42.4525 ms
99th percentile service time date_histogram_minute_agg 42.1739 44.3545 44.2888 ms
100th percentile service time date_histogram_minute_agg 42.3934 49.8299 50.052 ms
50th percentile latency range-auto-date-histo 2.14E+06 2.19E+06 2.18E+06 ms
90th percentile latency range-auto-date-histo 2.48E+06 2.54E+06 2.53E+06 ms
99th percentile latency range-auto-date-histo 2.55E+06 2.61E+06 2.61E+06 ms
100th percentile latency range-auto-date-histo 2.56E+06 2.62E+06 2.61E+06 ms
50th percentile service time range-auto-date-histo 9014.39 9218.99 9183.65 ms
90th percentile service time range-auto-date-histo 9167.03 9306.23 9320.49 ms
99th percentile service time range-auto-date-histo 9264.83 9410.94 9433.82 ms
100th percentile service time range-auto-date-histo 9321.41 9472.17 9531.95 ms

Latency and service time ONLY INCREASES WITH OS 2.19 index used with OS 3.0 server

Metric Name Operation/Query name OS 2.19 OS 3.0 OS 3.0 * OS 2.19 Unit
50th percentile latency range-auto-date-histo-with-metrics 5.53E+06 5.53E+06 5.71E+06 ms
90th percentile latency range-auto-date-histo-with-metrics 6.40E+06 6.40E+06 6.61E+06 ms
99th percentile latency range-auto-date-histo-with-metrics 6.59E+06 6.60E+06 6.81E+06 ms
100th percentile latency range-auto-date-histo-with-metrics 6.61E+06 6.62E+06 6.83E+06 ms
50th percentile service time range-auto-date-histo-with-metrics 22510.1 22540.7 23269.2 ms
90th percentile service time range-auto-date-histo-with-metrics 22722.3 22712.5 23621.4 ms
99th percentile service time range-auto-date-histo-with-metrics 22990.4 22860.9 23834.2 ms
100th percentile service time range-auto-date-histo-with-metrics 23000.1 22918.4 23857.7 ms

Related component

Search:Aggregations

To Reproduce

Run the above mentioned operations with Big5 workload using OS 3.0 server and compare with indices created in OS 2.19 v/s OS 3.0

Expected behavior

All operations should have the same latency or lesser with OS 3.0 irrespective of the underlying index version.

Additional Details

Meta

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Projects
Status: 🆕 New
Development

No branches or pull requests

1 participant