chore(ci): optimize size/usage of pre-built CI images

[tobz]

Summary

As stated in the PR title, really.

Every CI job starts off by pulling the configured container image, which means bigger images on a cold runner node (image not cached) will take longer before the actual job itself can start.

Change Type

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

How did you test this PR?

The images build cleanly in CI, and seem to work for dependent jobs. 🤞🏻

References

DADP-11

[pr-commenter]

Regression Detector (Agent Data Plane)

Regression Detector Results

Run ID: 4245fe56-906c-4b72-8aad-2ddf7e2dae55

Baseline: 4dd143e
Comparison: 136a97b
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_cpu	% cpu utilization	+0.85	[-4.12, +5.82]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_logs_5mb_throughput	ingress throughput	+0.00	[-0.13, +0.13]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_logs_5mb_memory	memory utilization	-4.26	[-4.73, -3.80]	1	(metrics) (profiles) (logs)

Fine details of change detection per experiment

perf	experiment	goal	Δ mean %	Δ mean % CI	trials	links
➖	dsd_uds_100mb_3k_contexts_cpu	% cpu utilization	+2.29	[-3.99, +8.57]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_traces_ottl_filtering_5mb_cpu	% cpu utilization	+0.87	[-1.50, +3.24]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_logs_5mb_cpu	% cpu utilization	+0.85	[-4.12, +5.82]	1	(metrics) (profiles) (logs)
➖	dsd_uds_512kb_3k_contexts_cpu	% cpu utilization	+0.68	[-56.27, +57.63]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_metrics_5mb_memory	memory utilization	+0.64	[+0.48, +0.79]	1	(metrics) (profiles) (logs)
➖	dsd_uds_100mb_3k_contexts_memory	memory utilization	+0.63	[+0.47, +0.78]	1	(metrics) (profiles) (logs)
➖	dsd_uds_500mb_3k_contexts_memory	memory utilization	+0.57	[+0.41, +0.72]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_traces_ottl_transform_5mb_memory	memory utilization	+0.35	[+0.19, +0.51]	1	(metrics) (profiles) (logs)
➖	dsd_uds_500mb_3k_contexts_cpu	% cpu utilization	+0.25	[-1.15, +1.66]	1	(metrics) (profiles) (logs)
➖	quality_gates_rss_dsd_medium	memory utilization	+0.24	[+0.07, +0.41]	1	(metrics) (profiles) (logs)
➖	quality_gates_rss_dsd_ultraheavy	memory utilization	+0.13	[+0.00, +0.25]	1	(metrics) (profiles) (logs)
➖	quality_gates_rss_idle	memory utilization	+0.10	[+0.06, +0.14]	1	(metrics) (profiles) (logs)
➖	dsd_uds_512kb_3k_contexts_memory	memory utilization	+0.09	[-0.05, +0.23]	1	(metrics) (profiles) (logs)
➖	dsd_uds_10mb_3k_contexts_memory	memory utilization	+0.07	[-0.09, +0.22]	1	(metrics) (profiles) (logs)
➖	dsd_uds_1mb_3k_contexts_memory	memory utilization	+0.06	[-0.08, +0.21]	1	(metrics) (profiles) (logs)
➖	quality_gates_rss_dsd_low	memory utilization	+0.06	[-0.11, +0.23]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_traces_5mb_throughput	ingress throughput	+0.01	[-0.07, +0.08]	1	(metrics) (profiles) (logs)
➖	dsd_uds_100mb_3k_contexts_throughput	ingress throughput	+0.00	[-0.04, +0.05]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_logs_5mb_throughput	ingress throughput	+0.00	[-0.13, +0.13]	1	(metrics) (profiles) (logs)
➖	dsd_uds_512kb_3k_contexts_throughput	ingress throughput	+0.00	[-0.05, +0.05]	1	(metrics) (profiles) (logs)
➖	dsd_uds_10mb_3k_contexts_throughput	ingress throughput	+0.00	[-0.14, +0.14]	1	(metrics) (profiles) (logs)
➖	dsd_uds_1mb_3k_contexts_throughput	ingress throughput	-0.00	[-0.06, +0.05]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_traces_ottl_filtering_5mb_memory	memory utilization	-0.00	[-0.25, +0.25]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_traces_ottl_transform_5mb_cpu	% cpu utilization	-0.02	[-2.10, +2.06]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_metrics_5mb_throughput	ingress throughput	-0.03	[-0.20, +0.13]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_traces_ottl_filtering_5mb_throughput	ingress throughput	-0.10	[-0.17, -0.02]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_metrics_5mb_cpu	% cpu utilization	-0.13	[-8.05, +7.79]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_traces_ottl_transform_5mb_throughput	ingress throughput	-0.13	[-0.21, -0.06]	1	(metrics) (profiles) (logs)
➖	dsd_uds_1mb_3k_contexts_cpu	% cpu utilization	-0.17	[-54.81, +54.48]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_traces_5mb_memory	memory utilization	-0.23	[-0.39, -0.06]	1	(metrics) (profiles) (logs)
➖	quality_gates_rss_dsd_heavy	memory utilization	-0.24	[-0.36, -0.11]	1	(metrics) (profiles) (logs)
➖	dsd_uds_500mb_3k_contexts_throughput	ingress throughput	-0.87	[-0.99, -0.74]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_traces_5mb_cpu	% cpu utilization	-2.22	[-4.37, -0.07]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_logs_5mb_memory	memory utilization	-4.26	[-4.73, -3.80]	1	(metrics) (profiles) (logs)
➖	dsd_uds_10mb_3k_contexts_cpu	% cpu utilization	-7.71	[-37.65, +22.24]	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	120.58MiB ≤ 140MiB	(metrics) (profiles) (logs)
✅	quality_gates_rss_dsd_low	memory_usage	10/10	39.77MiB ≤ 50MiB	(metrics) (profiles) (logs)
✅	quality_gates_rss_dsd_medium	memory_usage	10/10	60.59MiB ≤ 75MiB	(metrics) (profiles) (logs)
✅	quality_gates_rss_dsd_ultraheavy	memory_usage	10/10	176.86MiB ≤ 200MiB	(metrics) (profiles) (logs)
✅	quality_gates_rss_idle	memory_usage	10/10	27.00MiB ≤ 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:

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".

[pr-commenter]

Binary Size Analysis (Agent Data Plane)

Target: 4dd143e (baseline) vs 136a97b (comparison) diff
Analysis Type: Stripped binaries (debug symbols excluded)
Baseline Size: 37.09 MiB
Comparison Size: 37.09 MiB
Size Change: +0 B (+0.00%)
Pass/Fail Threshold: +5%
Result: PASSED ✅

Changes by Module

Module	File Size	Symbols
anon.4f8fd67d74ae1f1600187cfeb0121be9.1.llvm.6126630356860252161	+129 B	1
anon.4f8fd67d74ae1f1600187cfeb0121be9.1.llvm.4033737582525021100	-129 B	1
anon.4f8fd67d74ae1f1600187cfeb0121be9.4.llvm.6126630356860252161	+114 B	1
anon.4f8fd67d74ae1f1600187cfeb0121be9.4.llvm.4033737582525021100	-114 B	1
anon.4f8fd67d74ae1f1600187cfeb0121be9.3.llvm.6126630356860252161	+108 B	1
anon.4f8fd67d74ae1f1600187cfeb0121be9.3.llvm.4033737582525021100	-108 B	1
anon.4f8fd67d74ae1f1600187cfeb0121be9.0.llvm.6126630356860252161	+96 B	1
anon.4f8fd67d74ae1f1600187cfeb0121be9.0.llvm.4033737582525021100	-96 B	1
anon.4f8fd67d74ae1f1600187cfeb0121be9.2.llvm.6126630356860252161	+94 B	1
anon.4f8fd67d74ae1f1600187cfeb0121be9.2.llvm.4033737582525021100	-94 B	1

Detailed Symbol Changes

    FILE SIZE        VM SIZE    
 --------------  -------------- 
  [NEW]    +129  [NEW]     +40    anon.4f8fd67d74ae1f1600187cfeb0121be9.1.llvm.6126630356860252161
  [NEW]    +114  [NEW]     +25    anon.4f8fd67d74ae1f1600187cfeb0121be9.4.llvm.6126630356860252161
  [NEW]    +108  [NEW]     +19    anon.4f8fd67d74ae1f1600187cfeb0121be9.3.llvm.6126630356860252161
  [NEW]     +96  [NEW]      +7    anon.4f8fd67d74ae1f1600187cfeb0121be9.0.llvm.6126630356860252161
  [NEW]     +94  [NEW]      +5    anon.4f8fd67d74ae1f1600187cfeb0121be9.2.llvm.6126630356860252161
  [DEL]     -94  [DEL]      -5    anon.4f8fd67d74ae1f1600187cfeb0121be9.2.llvm.4033737582525021100
  [DEL]     -96  [DEL]      -7    anon.4f8fd67d74ae1f1600187cfeb0121be9.0.llvm.4033737582525021100
  [DEL]    -108  [DEL]     -19    anon.4f8fd67d74ae1f1600187cfeb0121be9.3.llvm.4033737582525021100
  [DEL]    -114  [DEL]     -25    anon.4f8fd67d74ae1f1600187cfeb0121be9.4.llvm.4033737582525021100
  [DEL]    -129  [DEL]     -40    anon.4f8fd67d74ae1f1600187cfeb0121be9.1.llvm.4033737582525021100
  [ = ]       0  [ = ]       0    TOTAL

[andrewqian2001datadog]

The script /install-awscli.sh only installs the cli for x86, do we want other architectures as well?

install_awscli() {
    local url="https://awscli.amazonaws.com/awscli-exe-linux-x86_64.zip"

[tobz]

This is true, but we only run the jobs in benchmark.yml on AMD64 runners.

[tobz]

Although saying that makes me realize... we still technically build the SMP CI image for linux/arm64, even if we never end up using that version. Kind of a footgun... I'll add it as a follow-up issue to address just to avoid someone else stumbling over that in the future.