enhancement(ground-truth): add parallel test runner

[thieman]

Summary

• Replaces all individual sequential Makefile targets with a single test-correctness target that discovers and runs all correctness test cases in parallel
• Adds run-all subcommand to ground-truth that mirrors how panoramic manages integration tests: uses stream::iter + buffer_unordered for semaphore-free concurrency limiting, with a default parallelism of 4
• Adds run subcommand to ground-truth for single test case invocation (replaces old bare positional arg); updates GitLab CI jobs accordingly
• New test cases are automatically picked up without any Makefile or CI changes

Details

ground-truth previously took a single config file path as a positional argument. It now has two subcommands:

• run <config.yaml> — runs a single test case (existing behavior, used by GitLab CI jobs)
• run-all -d <dir> [-p <parallelism>] — discovers all config.yaml files in subdirectories and runs them in parallel (default parallelism: 4)

The Makefile gains:

• make test-correctness — runs the full suite via run-all (parallelism overridable with GROUND_TRUTH_PARALLELISM=N)
• make test-correctness-case CASE=<name> — escape hatch for a single named case

Motivated by tobz's note in #1344.

Test plan

• cargo build --package ground-truth compiles cleanly
• make test-correctness runs all cases in parallel
• make test-correctness GROUND_TRUTH_PARALLELISM=8 overrides parallelism
• make test-correctness-case CASE=dsd-plain runs a single case
• Adding a new test case directory is auto-discovered without Makefile/CI changes

🤖 Generated with Claude Code

[pr-commenter]

Binary Size Analysis (Agent Data Plane)

Target: 644871f (baseline) vs cde8485 (comparison) diff
Analysis Type: Stripped binaries (debug symbols excluded)
Baseline Size: 26.60 MiB
Comparison Size: 26.60 MiB
Size Change: +0 B (+0.00%)
Pass/Fail Threshold: +5%
Result: PASSED ✅

Changes by Module

Module	File Size	Symbols

Detailed Symbol Changes

    FILE SIZE        VM SIZE    
 --------------  -------------- 
  [ = ]       0  [ = ]       0    TOTAL

[pr-commenter]

Regression Detector (Agent Data Plane)

Regression Detector Results

Run ID: c0eff7b9-8d65-435f-9715-61910be36b11

Baseline: 644871f
Comparison: cde8485
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_throughput	ingress throughput	+0.01	[-0.12, +0.13]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_logs_5mb_cpu	% cpu utilization	-1.94	[-6.74, +2.86]	1	(metrics) (profiles) (logs)
✅	otlp_ingest_logs_5mb_memory	memory utilization	-5.04	[-5.48, -4.60]	1	(metrics) (profiles) (logs)

Fine details of change detection per experiment

perf	experiment	goal	Δ mean %	Δ mean % CI	trials	links
➖	otlp_ingest_traces_5mb_cpu	% cpu utilization	+1.76	[-0.47, +3.99]	1	(metrics) (profiles) (logs)
➖	dsd_uds_512kb_3k_contexts_cpu	% cpu utilization	+0.79	[-56.68, +58.27]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_traces_ottl_filtering_5mb_memory	memory utilization	+0.58	[+0.24, +0.92]	1	(metrics) (profiles) (logs)
➖	dsd_uds_500mb_3k_contexts_throughput	ingress throughput	+0.57	[+0.44, +0.70]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_traces_ottl_filtering_5mb_cpu	% cpu utilization	+0.51	[-1.85, +2.86]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_traces_ottl_transform_5mb_memory	memory utilization	+0.42	[+0.17, +0.66]	1	(metrics) (profiles) (logs)
➖	dsd_uds_500mb_3k_contexts_cpu	% cpu utilization	+0.37	[-1.13, +1.86]	1	(metrics) (profiles) (logs)
➖	quality_gates_rss_dsd_heavy	memory utilization	+0.26	[+0.12, +0.39]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_metrics_5mb_cpu	% cpu utilization	+0.26	[-5.50, +6.02]	1	(metrics) (profiles) (logs)
➖	dsd_uds_1mb_3k_contexts_memory	memory utilization	+0.13	[-0.04, +0.29]	1	(metrics) (profiles) (logs)
➖	dsd_uds_500mb_3k_contexts_memory	memory utilization	+0.12	[-0.05, +0.29]	1	(metrics) (profiles) (logs)
➖	dsd_uds_512kb_3k_contexts_memory	memory utilization	+0.09	[-0.08, +0.26]	1	(metrics) (profiles) (logs)
➖	dsd_uds_100mb_3k_contexts_throughput	ingress throughput	+0.01	[-0.04, +0.05]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_logs_5mb_throughput	ingress throughput	+0.01	[-0.12, +0.13]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_traces_5mb_memory	memory utilization	+0.01	[-0.24, +0.26]	1	(metrics) (profiles) (logs)
➖	dsd_uds_1mb_3k_contexts_throughput	ingress throughput	+0.00	[-0.05, +0.06]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_traces_ottl_filtering_5mb_throughput	ingress throughput	-0.00	[-0.02, +0.02]	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)
➖	otlp_ingest_metrics_5mb_throughput	ingress throughput	-0.01	[-0.13, +0.12]	1	(metrics) (profiles) (logs)
➖	dsd_uds_10mb_3k_contexts_throughput	ingress throughput	-0.01	[-0.15, +0.14]	1	(metrics) (profiles) (logs)
➖	dsd_uds_1mb_3k_contexts_cpu	% cpu utilization	-0.01	[-53.22, +53.20]	1	(metrics) (profiles) (logs)
➖	quality_gates_rss_dsd_ultraheavy	memory utilization	-0.02	[-0.14, +0.11]	1	(metrics) (profiles) (logs)
➖	quality_gates_rss_dsd_medium	memory utilization	-0.06	[-0.25, +0.13]	1	(metrics) (profiles) (logs)
➖	quality_gates_rss_dsd_low	memory utilization	-0.06	[-0.25, +0.12]	1	(metrics) (profiles) (logs)
➖	dsd_uds_100mb_3k_contexts_memory	memory utilization	-0.07	[-0.25, +0.11]	1	(metrics) (profiles) (logs)
➖	quality_gates_rss_idle	memory utilization	-0.15	[-0.18, -0.12]	1	(metrics) (profiles) (logs)
➖	dsd_uds_10mb_3k_contexts_memory	memory utilization	-0.40	[-0.58, -0.22]	1	(metrics) (profiles) (logs)
➖	dsd_uds_100mb_3k_contexts_cpu	% cpu utilization	-1.00	[-7.43, +5.42]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_logs_5mb_cpu	% cpu utilization	-1.94	[-6.74, +2.86]	1	(metrics) (profiles) (logs)
➖	dsd_uds_10mb_3k_contexts_cpu	% cpu utilization	-3.08	[-33.70, +27.54]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_traces_ottl_transform_5mb_cpu	% cpu utilization	-3.72	[-5.89, -1.54]	1	(metrics) (profiles) (logs)
➖	otlp_ingest_metrics_5mb_memory	memory utilization	-4.40	[-4.60, -4.20]	1	(metrics) (profiles) (logs)
✅	otlp_ingest_logs_5mb_memory	memory utilization	-5.04	[-5.48, -4.60]	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	113.74MiB ≤ 140MiB	(metrics) (profiles) (logs)
✅	quality_gates_rss_dsd_low	memory_usage	10/10	34.61MiB ≤ 50MiB	(metrics) (profiles) (logs)
✅	quality_gates_rss_dsd_medium	memory_usage	10/10	54.42MiB ≤ 75MiB	(metrics) (profiles) (logs)
✅	quality_gates_rss_dsd_ultraheavy	memory_usage	10/10	171.56MiB ≤ 200MiB	(metrics) (profiles) (logs)
✅	quality_gates_rss_idle	memory_usage	10/10	21.72MiB ≤ 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".

[tobz]

I like that this makes things go a little faster, but the output is a bit hard to comb through because it's all interleaved without any identifying information about which log lines belong to which test case, etc.

I do think we really ought to strive to model this more on how panoramic does it, where it's got a nice TUI mode that hides a lot of the details at info level but then has a non-TUI mode for CI. It would also be interesting to explore just merging the two together, so that panoramic runs everything, and we just have two classes of tests it supports (and the ability to add more in the future), etc...

[thieman]

👍 I'll look more closely at the panoramic model

[thieman]

[Claude Sonnet 4.6] Closing in favor of a fuller merge of ground-truth into panoramic.

[thieman]

Tackled this here #1379