test(smp): add initial Agent+ADP tagfilter benchmark harness by rayz · Pull Request #1327 · DataDog/saluki

rayz · 2026-04-08T19:58:04Z

Summary

do not merge. testing converged image + tag filtering

the job to look for is run-benchmarks-agent-adp

Change Type

Bug fix
New feature
Non-functional (chore, refactoring, docs)
Performance

How did you test this PR?

References

pr-commenter · 2026-04-08T20:02:49Z

Binary Size Analysis (Agent Data Plane)

Target: 1fec9ae (baseline) vs 67343e8 (comparison) diff
Analysis Type: Stripped binaries (debug symbols excluded)
Baseline Size: 26.55 MiB
Comparison Size: 26.56 MiB
Size Change: +4.43 KiB (+0.02%)
Pass/Fail Threshold: +5%
Result: PASSED ✅

Changes by Module

Module	File Size	Symbols
`[Unmapped]`	+2.04 KiB	1
`agent_data_plane::cli::run`	+1.18 KiB	4
`serde_core`	+614 B	1
`saluki_env::helpers::remote_agent`	+595 B	1
`saluki_env::workload::providers`	+41 B	1
`agent_data_plane::components::tag_filterlist`	-35 B	1
`[sections]`	+30 B	8
`core`	-19 B	7
`aho_corasick`	+12 B	1
`std`	-4 B	1
`figment`	-2 B	1
`[LOAD #3 [RW]]`	+0 B	1

Detailed Symbol Changes

    FILE SIZE        VM SIZE    
 --------------  -------------- 
   +62% +2.04Ki  [ = ]       0    [Unmapped]
  [NEW]    +680  [NEW]    +527    core::ptr::drop_in_place<saluki_env::helpers::remote_agent::client::RemoteAgentClientConfiguration>::h27687f06e772f244
  [NEW]    +614  [NEW]    +527    serde_core::de::SeqAccess::next_element::hbed77297a98bc7ac
  +1.8%    +595  +1.9%    +595    saluki_env::helpers::remote_agent::client::RemoteAgentClient::from_configuration::_{{closure}}::hf89ea40936fb5711
  [NEW]    +429  [NEW]    +288    agent_data_plane::cli::run::handle_run_command::_{{closure}}::__CALLSITE::META::h5ecf203af00aacc6
  +0.6%    +390  +0.6%    +390    agent_data_plane::cli::run::handle_run_command::_{{closure}}::hbff4cb15be99fca3
  +0.2%    +208  +0.2%    +208    agent_data_plane::cli::run::create_topology::_{{closure}}::h5c132aa500e2a433
  [NEW]    +184  [NEW]     +48    agent_data_plane::cli::run::handle_run_command::_{{closure}}::__CALLSITE::h887b740782d36205
  +0.0%     +84  +0.0%     +84    [section .gcc_except_table]
  +1.0%     +72  +1.0%     +72    [section .rela.dyn]
  +0.0%     +47  +0.0%     +47    [section .text]
  +0.1%     +41  +0.1%     +41    saluki_env::workload::providers::remote_agent::RemoteAgentWorkloadProvider::from_configuration::_{{closure}}::hfc80d52378eff90b
  +0.5%     +32  +0.5%     +32    [section .data.rel.ro]
  +0.0%     +24  [ = ]       0    [section .symtab]
  +0.0%     +19  [ = ]       0    [section .strtab]
  +0.1%     +14  +0.2%     +14    [8 Others]
  -3.6%     -19  -6.1%     -19    core::ptr::drop_in_place<saluki_env::workload::collectors::remote_agent::tagger::RemoteAgentTaggerMetadataCollector::from_configuration::{{closure}}>::hb8b93f35d34f56ca
  [ = ]       0 -50.0%     -32    [LOAD #3 [RW]]
  -1.5%     -35  -1.7%     -35    _<agent_data_plane::components::tag_filterlist::TagFilterlistConfiguration as saluki_core::components::transforms::builder::TransformBuilder>::build::_{{closure}}::h4e972d76f411f997
  -5.8%    -256  -5.8%    -256    [section .rodata]
  [DEL]    -680  [DEL]    -527    core::ptr::drop_in_place<saluki_env::helpers::remote_agent::client::RemoteAgentClientConfiguration>::hf4e807b2a3c33f1c
  +0.0% +4.43Ki  +0.0% +1.96Ki    TOTAL

pr-commenter · 2026-04-08T20:20:25Z

Regression Detector (Agent Data Plane)

Regression Detector Results

Run ID: 51905a2f-fece-4ae9-8016-2979826a939f

Baseline: 1fec9ae
Comparison: 67343e8
Diff

Optimization Goals: ✅ No significant changes detected

Experiments ignored for regressions

Regressions in experiments with settings containing erratic: true are ignored.

perf	experiment	goal	Δ mean %	Δ mean % CI	trials	links
➖	otlp_ingest_logs_5mb_memory	memory utilization	+2.16	[+1.46, +2.86]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_logs_5mb_cpu	% cpu utilization	+0.01	[-4.76, +4.79]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_logs_5mb_throughput	ingress throughput	+0.00	[-0.12, +0.12]	1	(metrics) (profiles) (logs)

Fine details of change detection per experiment

perf	experiment	goal	Δ mean %	Δ mean % CI	trials	links
➖	dsd_uds_512kb_3k_contexts_cpu	% cpu utilization	+6.75	[-52.59, +66.09]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_logs_5mb_memory	memory utilization	+2.16	[+1.46, +2.86]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_traces_ottl_filtering_5mb_cpu	% cpu utilization	+0.76	[-1.73, +3.25]	1	(metrics) (profiles) (logs)
➖	dsd_uds_100mb_3k_contexts_cpu	% cpu utilization	+0.53	[-5.56, +6.62]	1	(metrics) (profiles) (logs)
➖	quality_gates_rss_dsd_ultraheavy	memory utilization	+0.49	[+0.37, +0.61]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_traces_ottl_filtering_5mb_memory	memory utilization	+0.47	[+0.14, +0.80]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_logs_5mb_cpu	% cpu utilization	+0.01	[-4.76, +4.79]	1	(metrics) (profiles) (logs)
➖	dsd_uds_100mb_3k_contexts_throughput	ingress throughput	+0.01	[-0.01, +0.03]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_traces_ottl_filtering_5mb_throughput	ingress throughput	+0.01	[-0.02, +0.03]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_metrics_5mb_throughput	ingress throughput	+0.00	[-0.12, +0.13]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_logs_5mb_throughput	ingress throughput	+0.00	[-0.12, +0.12]	1	(metrics) (profiles) (logs)
➖	dsd_uds_1mb_3k_contexts_throughput	ingress throughput	-0.00	[-0.06, +0.06]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_traces_5mb_throughput	ingress throughput	-0.00	[-0.02, +0.02]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_traces_ottl_transform_5mb_throughput	ingress throughput	-0.00	[-0.02, +0.02]	1	(metrics) (profiles) (logs)
➖	dsd_uds_512kb_3k_contexts_throughput	ingress throughput	-0.00	[-0.06, +0.05]	1	(metrics) (profiles) (logs)
➖	quality_gates_rss_dsd_medium	memory utilization	-0.00	[-0.19, +0.18]	1	(metrics) (profiles) (logs)
➖	dsd_uds_10mb_3k_contexts_throughput	ingress throughput	-0.01	[-0.12, +0.11]	1	(metrics) (profiles) (logs)
➖	dsd_uds_100mb_3k_contexts_memory	memory utilization	-0.02	[-0.20, +0.17]	1	(metrics) (profiles) (logs)
➖	dsd_uds_500mb_3k_contexts_memory	memory utilization	-0.02	[-0.18, +0.14]	1	(metrics) (profiles) (logs)
➖	dsd_uds_512kb_3k_contexts_memory	memory utilization	-0.03	[-0.20, +0.14]	1	(metrics) (profiles) (logs)
➖	quality_gates_rss_idle	memory utilization	-0.08	[-0.11, -0.05]	1	(metrics) (profiles) (logs)
➖	dsd_uds_10mb_3k_contexts_memory	memory utilization	-0.09	[-0.27, +0.09]	1	(metrics) (profiles) (logs)
➖	dsd_uds_1mb_3k_contexts_memory	memory utilization	-0.13	[-0.30, +0.04]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_metrics_5mb_memory	memory utilization	-0.20	[-0.38, -0.02]	1	(metrics) (profiles) (logs)
➖	quality_gates_rss_dsd_low	memory utilization	-0.25	[-0.44, -0.06]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_traces_5mb_memory	memory utilization	-0.32	[-0.56, -0.08]	1	(metrics) (profiles) (logs)
➖	quality_gates_rss_dsd_heavy	memory utilization	-0.38	[-0.52, -0.25]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_traces_ottl_transform_5mb_memory	memory utilization	-0.50	[-0.75, -0.25]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_metrics_5mb_cpu	% cpu utilization	-0.64	[-7.31, +6.03]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_traces_5mb_cpu	% cpu utilization	-0.72	[-2.90, +1.46]	1	(metrics) (profiles) (logs)
➖	dsd_uds_500mb_3k_contexts_cpu	% cpu utilization	-0.75	[-2.17, +0.67]	1	(metrics) (profiles) (logs)
➖	dsd_uds_500mb_3k_contexts_throughput	ingress throughput	-2.29	[-2.42, -2.16]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_traces_ottl_transform_5mb_cpu	% cpu utilization	-3.59	[-5.80, -1.38]	1	(metrics) (profiles) (logs)
➖	dsd_uds_1mb_3k_contexts_cpu	% cpu utilization	-7.42	[-58.91, +44.07]	1	(metrics) (profiles) (logs)
➖	dsd_uds_10mb_3k_contexts_cpu	% cpu utilization	-7.88	[-37.59, +21.83]	1	(metrics) (profiles) (logs)

Bounds Checks: ✅ Passed

perf	experiment	bounds_check_name	replicates_passed	observed_value	links
✅	quality_gates_rss_dsd_heavy	memory_usage	10/10	115.33MiB ≤ 140MiB	(metrics) (profiles) (logs)
✅	quality_gates_rss_dsd_low	memory_usage	10/10	34.41MiB ≤ 50MiB	(metrics) (profiles) (logs)
✅	quality_gates_rss_dsd_medium	memory_usage	10/10	56.42MiB ≤ 75MiB	(metrics) (profiles) (logs)
✅	quality_gates_rss_dsd_ultraheavy	memory_usage	10/10	169.18MiB ≤ 200MiB	(metrics) (profiles) (logs)
✅	quality_gates_rss_idle	memory_usage	10/10	21.39MiB ≤ 40MiB	(metrics) (profiles) (logs)

Explanation

Confidence level: 90.00%
Effect size tolerance: |Δ mean %| ≥ 5.00%

Performance changes are noted in the perf column of each table:

✅ = significantly better comparison variant performance
❌ = significantly worse comparison variant performance
➖ = no significant change in performance

A regression test is an A/B test of target performance in a repeatable rig, where "performance" is measured as "comparison variant minus baseline variant" for an optimization goal (e.g., ingress throughput). Due to intrinsic variability in measuring that goal, we can only estimate its mean value for each experiment; we report uncertainty in that value as a 90.00% confidence interval denoted "Δ mean % CI".

For each experiment, we decide whether a change in performance is a "regression" -- a change worth investigating further -- if all of the following criteria are true:

Its estimated |Δ mean %| ≥ 5.00%, indicating the change is big enough to merit a closer look.
Its 90.00% confidence interval "Δ mean % CI" does not contain zero, indicating that if our statistical model is accurate, there is at least a 90.00% chance there is a difference in performance between baseline and comparison variants.
Its configuration does not mark it "erratic".

pr-commenter · 2026-04-08T21:37:15Z

Regression Detector (Agent + ADP Tag Filtering)

Regression Detector Results

Run ID: 64471026-1b83-4c86-956d-f0b2048f1431

Baseline: d88bacd9
Comparison: 67343e8
Diff

Optimization Goals: ✅ Improvement(s) detected

perf	experiment	goal	Δ mean %	Δ mean % CI	trials	links
✅	tagfilter_10000metrics	% cpu utilization	-21.90	[-43.41, -0.39]	1	bounds checks dashboard
✅	tagfilter_100metrics	% cpu utilization	-27.68	[-44.28, -11.08]	1	bounds checks dashboard
✅	tagfilter_0metrics	% cpu utilization	-28.86	[-47.05, -10.67]	1	bounds checks dashboard
✅	tagfilter_500metrics	% cpu utilization	-33.01	[-55.99, -10.04]	1	bounds checks dashboard

Fine details of change detection per experiment

perf	experiment	goal	Δ mean %	Δ mean % CI	trials	links
➖	tagfilter_500metrics_notraffic	% cpu utilization	-4.22	[-88.72, +80.27]	1	bounds checks dashboard
➖	tagfilter_15000metrics	% cpu utilization	-10.49	[-33.60, +12.62]	1	bounds checks dashboard
✅	tagfilter_10000metrics	% cpu utilization	-21.90	[-43.41, -0.39]	1	bounds checks dashboard
✅	tagfilter_100metrics	% cpu utilization	-27.68	[-44.28, -11.08]	1	bounds checks dashboard
✅	tagfilter_0metrics	% cpu utilization	-28.86	[-47.05, -10.67]	1	bounds checks dashboard
✅	tagfilter_500metrics	% cpu utilization	-33.01	[-55.99, -10.04]	1	bounds checks dashboard

Bounds Checks: ❌ Failed

perf	experiment	bounds_check_name	replicates_passed	observed_value	links
✅	tagfilter_0metrics	memory_usage	10/10	450.10MiB ≤ 900MiB	bounds checks dashboard
❌	tagfilter_10000metrics	memory_usage	0/10	3.22GiB > 0.88GiB	bounds checks dashboard
✅	tagfilter_100metrics	memory_usage	10/10	509.99MiB ≤ 900MiB	bounds checks dashboard
❌	tagfilter_15000metrics	memory_usage	0/10	5.03GiB > 0.88GiB	bounds checks dashboard
✅	tagfilter_500metrics	memory_usage	10/10	631.71MiB ≤ 900MiB	bounds checks dashboard
✅	tagfilter_500metrics_notraffic	memory_usage	10/10	472.59MiB ≤ 900MiB	bounds checks dashboard

Explanation

Confidence level: 90.00%
Effect size tolerance: |Δ mean %| ≥ 5.00%

Performance changes are noted in the perf column of each table:

✅ = significantly better comparison variant performance
❌ = significantly worse comparison variant performance
➖ = no significant change in performance

A regression test is an A/B test of target performance in a repeatable rig, where "performance" is measured as "comparison variant minus baseline variant" for an optimization goal (e.g., ingress throughput). Due to intrinsic variability in measuring that goal, we can only estimate its mean value for each experiment; we report uncertainty in that value as a 90.00% confidence interval denoted "Δ mean % CI".

For each experiment, we decide whether a change in performance is a "regression" -- a change worth investigating further -- if all of the following criteria are true:

Its estimated |Δ mean %| ≥ 5.00%, indicating the change is big enough to merit a closer look.
Its 90.00% confidence interval "Δ mean % CI" does not contain zero, indicating that if our statistical model is accurate, there is at least a 90.00% chance there is a difference in performance between baseline and comparison variants.
Its configuration does not mark it "erratic".

+
+import argparse
+import copy
+import os


pr-commenter · 2026-04-20T16:03:39Z

Regression Detector (ADP Tag Filtering)

Regression Detector Results

Run ID: 7eb83243-6f87-46e3-8937-ed58febbf823

Baseline: 1fec9ae
Comparison: 67343e8
Diff

Optimization Goals: ✅ No significant changes detected

Fine details of change detection per experiment

perf	experiment	goal	Δ mean %	Δ mean % CI	trials	links
➖	tagfilter_15000metrics	% cpu utilization	+9.45	[-8.02, +26.93]	1	(metrics) (profiles) (logs)
➖	tagfilter_100metrics	% cpu utilization	-0.04	[-3.17, +3.09]	1	(metrics) (profiles) (logs)
➖	tagfilter_0metrics	% cpu utilization	-0.12	[-2.59, +2.35]	1	(metrics) (profiles) (logs)
➖	tagfilter_10000metrics	% cpu utilization	-0.18	[-14.38, +14.02]	1	(metrics) (profiles) (logs)
➖	tagfilter_500metrics	% cpu utilization	-5.10	[-11.98, +1.78]	1	(metrics) (profiles) (logs)
➖	tagfilter_500metrics_notraffic	% cpu utilization	-19.16	[-85.66, +47.33]	1	(metrics) (profiles) (logs)

Bounds Checks: ✅ Passed

perf	experiment	bounds_check_name	replicates_passed	observed_value	links
✅	tagfilter_0metrics	memory_usage	10/10	272.47MiB ≤ 900MiB	(metrics) (profiles) (logs)
✅	tagfilter_10000metrics	memory_usage	10/10	496.59MiB ≤ 900MiB	(metrics) (profiles) (logs)
✅	tagfilter_100metrics	memory_usage	10/10	300.18MiB ≤ 900MiB	(metrics) (profiles) (logs)
✅	tagfilter_15000metrics	memory_usage	10/10	608.21MiB ≤ 900MiB	(metrics) (profiles) (logs)
✅	tagfilter_500metrics	memory_usage	10/10	304.23MiB ≤ 900MiB	(metrics) (profiles) (logs)
✅	tagfilter_500metrics_notraffic	memory_usage	10/10	32.87MiB ≤ 900MiB	(metrics) (profiles) (logs)

Explanation

Confidence level: 90.00%
Effect size tolerance: |Δ mean %| ≥ 5.00%

Performance changes are noted in the perf column of each table:

✅ = significantly better comparison variant performance
❌ = significantly worse comparison variant performance
➖ = no significant change in performance

A regression test is an A/B test of target performance in a repeatable rig, where "performance" is measured as "comparison variant minus baseline variant" for an optimization goal (e.g., ingress throughput). Due to intrinsic variability in measuring that goal, we can only estimate its mean value for each experiment; we report uncertainty in that value as a 90.00% confidence interval denoted "Δ mean % CI".

For each experiment, we decide whether a change in performance is a "regression" -- a change worth investigating further -- if all of the following criteria are true:

Its estimated |Δ mean %| ≥ 5.00%, indicating the change is big enough to merit a closer look.
Its 90.00% confidence interval "Δ mean % CI" does not contain zero, indicating that if our statistical model is accurate, there is at least a 90.00% chance there is a difference in performance between baseline and comparison variants.
Its configuration does not mark it "erratic".

dd-octo-sts Bot added area/ci CI/CD, automated testing, etc. area/test All things testing: unit/integration, correctness, SMP regression, etc. labels Apr 8, 2026

rayz force-pushed the rayz/tagfilter-agent-adp-smp branch from 2fceea7 to ec9ca3e Compare April 9, 2026 20:12

rayz added 14 commits April 10, 2026 11:50

test(smp): add initial Agent+ADP tagfilter benchmark harness

4f1af31

Update lading

741a609

Update pipeline

27364c0

more ci changes

9510c1c

fix experiment

a2c49fc

add debug logs

08f209f

try to fix tests again

022c1df

try 500metrics experiment only

7ad28af

try fix again

29a52a9

scrape adp telemetry

0fc999a

fix config

fe7caa9

set contexts limit to 80k

a668cb2

Add more experiments

c64daae

Remove debug logging

1981c5f

rayz force-pushed the rayz/tagfilter-agent-adp-smp branch from ec9ca3e to 1981c5f Compare April 10, 2026 15:53

rayz added 6 commits April 10, 2026 11:57

Add 10k and 15k experiments

db3790b

Try enabling configstrema

93a64f3

Specify dogstatd_string_interner_size

089e16d

Use new agent image

3dcaedc

Add job for solo ADP filtering

90d74d2

try increase grpc max message size

67343e8

dd-octo-sts Bot added the area/config Configuration. label Apr 20, 2026

github-code-quality Bot found potential problems Apr 20, 2026

View reviewed changes

Comment thread test/smp/regression/adp-tagfilter/generate_experiments.py

import argparse

import copy

import os

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

test(smp): add initial Agent+ADP tagfilter benchmark harness#1327

test(smp): add initial Agent+ADP tagfilter benchmark harness#1327
rayz wants to merge 20 commits intomainfrom
rayz/tagfilter-agent-adp-smp

rayz commented Apr 8, 2026 •

edited

Loading

Uh oh!

pr-commenter Bot commented Apr 8, 2026 •

edited

Loading

Changes by Module

Detailed Symbol Changes

Uh oh!

pr-commenter Bot commented Apr 8, 2026 •

edited

Loading

Experiments ignored for regressions

Fine details of change detection per experiment

Bounds Checks: ✅ Passed

Explanation

Uh oh!

pr-commenter Bot commented Apr 8, 2026 •

edited

Loading

Fine details of change detection per experiment

Explanation

Uh oh!

pr-commenter Bot commented Apr 20, 2026

Fine details of change detection per experiment

Bounds Checks: ✅ Passed

Explanation

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

Conversation

rayz commented Apr 8, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Change Type

How did you test this PR?

References

Uh oh!

pr-commenter Bot commented Apr 8, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Binary Size Analysis (Agent Data Plane)

Changes by Module

Detailed Symbol Changes

Uh oh!

pr-commenter Bot commented Apr 8, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Regression Detector (Agent Data Plane)

Regression Detector Results

Optimization Goals: ✅ No significant changes detected

Experiments ignored for regressions

Fine details of change detection per experiment

Bounds Checks: ✅ Passed

Explanation

Uh oh!

pr-commenter Bot commented Apr 8, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Regression Detector (Agent + ADP Tag Filtering)

Regression Detector Results

Optimization Goals: ✅ Improvement(s) detected

Fine details of change detection per experiment

Bounds Checks: ❌ Failed

Explanation

Uh oh!

pr-commenter Bot commented Apr 20, 2026

Regression Detector (ADP Tag Filtering)

Regression Detector Results

Optimization Goals: ✅ No significant changes detected

Fine details of change detection per experiment

Bounds Checks: ✅ Passed

Explanation

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

rayz commented Apr 8, 2026 •

edited

Loading

pr-commenter Bot commented Apr 8, 2026 •

edited

Loading

pr-commenter Bot commented Apr 8, 2026 •

edited

Loading

pr-commenter Bot commented Apr 8, 2026 •

edited

Loading