enhancement(io): rework how framing interacts with I/O buffers

[tobz]

Summary

Work in progress.

Change Type

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

How did you test this PR?

References

[pr-commenter]

Regression Detector (Agent Data Plane)

Regression Detector Results

Run ID: a22c8a00-c443-4b6e-8215-6ceb2a42006d

Baseline: 2b02146
Comparison: cf6bd87
Diff

❌ Experiments with retried target crashes

This is a critical error. One or more replicates failed with a non-zero exit code. These replicates may have been retried. See Replicate Execution Details for more information.

• otlp_ingest_traces_5mb_cpu
• quality_gates_rss_dsd_low
• quality_gates_rss_dsd_ultraheavy
• otlp_ingest_metrics_5mb_throughput

Optimization Goals: ✅ Improvement(s) detected

perf	experiment	goal	Δ mean %	Δ mean % CI	trials	links
✅	dsd_uds_500mb_3k_contexts_throughput	ingress throughput	+8.83	[+8.70, +8.96]	1	(metrics) (profiles) (logs)
✅	dsd_uds_100mb_3k_contexts_cpu	% cpu utilization	-6.90	[-12.79, -1.01]	1	(metrics) (profiles) (logs)

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	+8.48	[+7.96, +9.00]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_logs_5mb_throughput	ingress throughput	-0.02	[-0.14, +0.11]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_logs_5mb_cpu	% cpu utilization	-0.16	[-5.46, +5.13]	1	(metrics) (profiles) (logs)

Fine details of change detection per experiment

perf	experiment	goal	Δ mean %	Δ mean % CI	trials	links
✅	dsd_uds_500mb_3k_contexts_throughput	ingress throughput	+8.83	[+8.70, +8.96]	1	(metrics) (profiles) (logs)
❌	otlp_ingest_logs_5mb_memory	memory utilization	+8.48	[+7.96, +9.00]	1	(metrics) (profiles) (logs)
➖	dsd_uds_1mb_3k_contexts_cpu	% cpu utilization	+6.15	[-49.32, +61.62]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_metrics_5mb_memory	memory utilization	+3.56	[+3.33, +3.80]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_traces_5mb_cpu	% cpu utilization	+3.43	[+1.24, +5.62]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_metrics_5mb_cpu	% cpu utilization	+2.74	[-3.24, +8.73]	1	(metrics) (profiles) (logs)
➖	dsd_uds_500mb_3k_contexts_memory	memory utilization	+1.21	[+1.02, +1.40]	1	(metrics) (profiles) (logs)
➖	quality_gates_rss_dsd_low	memory utilization	+0.74	[+0.57, +0.90]	1	(metrics) (profiles) (logs)
➖	dsd_uds_1mb_3k_contexts_memory	memory utilization	+0.55	[+0.36, +0.73]	1	(metrics) (profiles) (logs)
➖	dsd_uds_100mb_3k_contexts_memory	memory utilization	+0.19	[-0.01, +0.39]	1	(metrics) (profiles) (logs)
➖	quality_gates_rss_dsd_heavy	memory utilization	+0.16	[+0.03, +0.28]	1	(metrics) (profiles) (logs)
➖	dsd_uds_512kb_3k_contexts_memory	memory utilization	+0.03	[-0.15, +0.21]	1	(metrics) (profiles) (logs)
➖	dsd_uds_100mb_3k_contexts_throughput	ingress throughput	+0.02	[-0.03, +0.08]	1	(metrics) (profiles) (logs)
➖	dsd_uds_10mb_3k_contexts_throughput	ingress throughput	+0.01	[-0.15, +0.16]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_traces_5mb_throughput	ingress throughput	-0.00	[-0.02, +0.02]	1	(metrics) (profiles) (logs)
➖	dsd_uds_1mb_3k_contexts_throughput	ingress throughput	-0.00	[-0.06, +0.05]	1	(metrics) (profiles) (logs)
➖	dsd_uds_512kb_3k_contexts_throughput	ingress throughput	-0.00	[-0.06, +0.05]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_logs_5mb_throughput	ingress throughput	-0.02	[-0.14, +0.11]	1	(metrics) (profiles) (logs)
➖	quality_gates_rss_dsd_medium	memory utilization	-0.02	[-0.21, +0.16]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_metrics_5mb_throughput	ingress throughput	-0.04	[-0.17, +0.09]	1	(metrics) (profiles) (logs)
➖	quality_gates_rss_dsd_ultraheavy	memory utilization	-0.13	[-0.27, +0.01]	1	(metrics) (profiles) (logs)
➖	quality_gates_rss_idle	memory utilization	-0.13	[-0.18, -0.08]	1	(metrics) (profiles) (logs)
➖	dsd_uds_10mb_3k_contexts_memory	memory utilization	-0.14	[-0.34, +0.06]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_logs_5mb_cpu	% cpu utilization	-0.16	[-5.46, +5.13]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_traces_5mb_memory	memory utilization	-0.45	[-0.73, -0.16]	1	(metrics) (profiles) (logs)
➖	dsd_uds_500mb_3k_contexts_cpu	% cpu utilization	-0.61	[-1.94, +0.73]	1	(metrics) (profiles) (logs)
➖	dsd_uds_512kb_3k_contexts_cpu	% cpu utilization	-3.43	[-58.63, +51.77]	1	(metrics) (profiles) (logs)
➖	dsd_uds_10mb_3k_contexts_cpu	% cpu utilization	-6.35	[-36.11, +23.41]	1	(metrics) (profiles) (logs)
✅	dsd_uds_100mb_3k_contexts_cpu	% cpu utilization	-6.90	[-12.79, -1.01]	1	(metrics) (profiles) (logs)

Bounds Checks: ✅ Passed

perf	experiment	bounds_check_name	replicates_passed	links
✅	quality_gates_rss_dsd_heavy	memory_usage	10/10	(metrics) (profiles) (logs)
✅	quality_gates_rss_dsd_low	memory_usage	10/10	(metrics) (profiles) (logs)
✅	quality_gates_rss_dsd_medium	memory_usage	10/10	(metrics) (profiles) (logs)
✅	quality_gates_rss_dsd_ultraheavy	memory_usage	10/10	(metrics) (profiles) (logs)
✅	quality_gates_rss_idle	memory_usage	10/10	(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:

1. Its estimated |Δ mean %| ≥ 5.00%, indicating the change is big enough to merit a closer look.

2. 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.

3. Its configuration does not mark it "erratic".

Replicate Execution Details

We run multiple replicates for each experiment/variant. However, we allow replicates to be automatically retried if there are any failures, up to 8 times, at which point the replicate is marked dead and we are unable to run analysis for the entire experiment. We call each of these attempts at running replicates a replicate execution. This section lists all replicate executions that failed due to the target crashing or being oom killed.

Note: In the below tables we bucket failures by experiment, variant, and failure type. For each of these buckets we list out the replicate indexes that failed with an annotation signifying how many times said replicate failed with the given failure mode. In the below example the baseline variant of the experiment named experiment_with_failures had two replicates that failed by oom kills. Replicate 0, which failed 8 executions, and replicate 1 which failed 6 executions, all with the same failure mode.

Experiment	Variant	Replicates	Failure	Logs	Debug Dashboard
experiment_with_failures	baseline	0 (x8) 1 (x6)	Oom killed		Debug Dashboard

The debug dashboard links will take you to a debugging dashboard specifically designed to investigate replicate execution failures.

❌ Retried Normal Replicate Execution Failures (non-profiling)

Experiment	Variant	Replicates	Failure	Debug Dashboard
otlp_ingest_metrics_5mb_throughput	comparison	8	Failed to shutdown when requested	Debug Dashboard
otlp_ingest_traces_5mb_cpu	comparison	9	Failed to shutdown when requested	Debug Dashboard
quality_gates_rss_dsd_low	comparison	6	Failed to shutdown when requested	Debug Dashboard
quality_gates_rss_dsd_ultraheavy	comparison	7	Failed to shutdown when requested	Debug Dashboard

[lukesteensen]

I can't say I've really internalized how it works, but I think I get it and it seems reasonable 😅

[lukesteensen]

Does the remaining call here match with the use of chunk above? Seems like it could be wrong for non-contiguous buffers. It could be simpler to just keep track of how much we consume directly.

[tobz]

You're right on both fronts.

We never use non-contiguous buffers on the ingest side (BytesBuffer) so our usage of Buf here is somewhat loose here: this should be correct, but may not always be.

We could tie to more directly to BytesBuffer and revisit this generic approach later on if necessary, or we could do the bit you mentioned about tracking it directly... but I'm not sure I grok what you're envisioning there. Could you give a little more detail on that idea?

[lukesteensen]

Sorry, just realized I had a diff for this and totally forgot to respond 😞

Pushed to a branch here, feel free to take it or leave it. Claude added some tests and things you may or may not care about.

[tobz]

• enhancement(io): rework how framing interacts with I/O buffers #983 👈 (View in Graphite)
• main

This stack of pull requests is managed by Graphite. Learn more about stacking.

[pr-commenter]

Binary Size Analysis (Agent Data Plane)

Target: 2b02146 (baseline) vs cf6bd87 (comparison) diff
Analysis Type: Stripped binaries (debug symbols excluded)
Baseline Size: 26.87 MiB
Comparison Size: 26.86 MiB
Size Change: -10.45 KiB (-0.04%)
Pass/Fail Threshold: +5%
Result: PASSED ✅

Changes by Module

Module	File Size	Symbols
saluki_components::sources::dogstatsd	-8.61 KiB	6
saluki_io::deser::framing	-2.95 KiB	20
[Unmapped]	+2.71 KiB	1
bytes	-1.52 KiB	7
core	+1.04 KiB	38
hyper_rustls	-505 B	2
saluki_io::buf::vec	-192 B	1
[sections]	-189 B	7
saluki_common::task::instrument	-146 B	2
h2	+90 B	7
hyper	-81 B	2
saluki_components::common::datadog	-60 B	1
saluki_io::net::client	-60 B	1
unicode_segmentation	-48 B	1
hickory_proto	+43 B	11
__finalized.1738	-24 B	1
tokio	+14 B	32
saluki_components::sources::otlp	+5 B	1
std	-4 B	15
agent_data_plane::internal::remote_agent	-3 B	1

Detailed Symbol Changes

    FILE SIZE        VM SIZE    
 --------------  -------------- 
  [NEW] +17.6Ki  [NEW] +17.5Ki    _<saluki_common::task::instrument::InstrumentedTask<F> as core::future::future::Future>::poll::h667ef4daee0ab193
  [NEW] +9.34Ki  [NEW] +9.15Ki    _<saluki_components::sources::dogstatsd::DogStatsD as saluki_core::components::sources::Source>::run::_{{closure}}::h03e2f754767dbe2e
  [NEW] +5.13Ki  [NEW] +5.00Ki    <hickory_proto::rr::record_data::RData as core::clone::Clone>::clone.11464
  [NEW] +4.71Ki  [NEW] +4.59Ki    <rustls::error::Error as core::clone::Clone>::clone.11429
  [NEW] +3.83Ki  [NEW] +3.71Ki    <webpki::error::Error as core::fmt::Debug>::fmt.11626
  [NEW] +3.57Ki  [NEW] +3.45Ki    <rustls::error::Error as core::fmt::Debug>::fmt.11430
   +93% +2.71Ki  [ = ]       0    [Unmapped]
  [NEW] +1.79Ki  [NEW] +1.65Ki    <tonic::transport::channel::endpoint::Endpoint as core::clone::Clone>::clone.13194
  [NEW] +1.57Ki  [NEW] +1.50Ki    tokio::runtime::task::raw::poll::h8ae0139d00aaf8e1
  [NEW] +1.54Ki  [NEW] +1.46Ki    tokio::runtime::task::raw::poll::h141b155b261cedb3
  [DEL] -1.54Ki  [DEL] -1.46Ki    tokio::runtime::task::raw::poll::ha188c6dffca733f0
  [DEL] -1.57Ki  [DEL] -1.50Ki    tokio::runtime::task::raw::poll::h216e0a17c8efef6a
  [DEL] -1.79Ki  [DEL] -1.65Ki    <tonic::transport::channel::endpoint::Endpoint as core::clone::Clone>::clone.13208
  [DEL] -3.57Ki  [DEL] -3.45Ki    <rustls::error::Error as core::fmt::Debug>::fmt.11444
  [DEL] -3.83Ki  [DEL] -3.71Ki    <webpki::error::Error as core::fmt::Debug>::fmt.11640
  -0.3% -4.38Ki  -0.2% -2.05Ki    [229 Others]
  [DEL] -4.71Ki  [DEL] -4.59Ki    <rustls::error::Error as core::clone::Clone>::clone.11443
  [DEL] -5.13Ki  [DEL] -5.00Ki    <hickory_proto::rr::record_data::RData as core::clone::Clone>::clone.11478
 -26.5% -8.71Ki -26.6% -8.71Ki    saluki_components::sources::dogstatsd::drive_stream::_{{closure}}::h7a0b9c0eda95cb0d
  [DEL] -9.28Ki  [DEL] -9.09Ki    _<saluki_components::sources::dogstatsd::DogStatsD as saluki_core::components::sources::Source>::run::_{{closure}}::h7acbbb6388dcee1a
  [DEL] -17.8Ki  [DEL] -17.6Ki    _<saluki_common::task::instrument::InstrumentedTask<F> as core::future::future::Future>::poll::he2ecef1bad274487
  -0.0% -10.5Ki  -0.0% -10.8Ki    TOTAL