feat(ssi): last-TRUE-wins SSI matching

[iamluc]

What does this PR do?

Unify SSI matching on a single last-TRUE-wins matcher:

• Static targets (Helm / Operator / datadog.yaml) are reversed at construction so config first-wins is preserved.
• RC policies keep wire order (catch-all first, exceptions after) and are last-TRUE-wins as-is.
• Evaluation is static first, then RC, then SSI inject-all only when both are absent. A matched deny does not fall through.

SSI | static targets | RC          | Decision
off | —              | none        | nothing
off | —              | policies    | last matching policy, else nothing
on  | none           | none        | everything
on  | none           | policies    | last matching policy, else nothing
on  | present        | none        | first (as in the yaml) matching target, else nothing
on  | present        | policies    | first (as in the yaml) matching target, else last matching policy, else nothing

Rebased onto main after #55057 merged.

Motivation

RC is already last-wins on the wire. Using the same matcher for static targets (by reversing them) avoids two matching modes. Static targeting is evaluated before RC so a remote deny cannot override an explicit local target.

Describe how you validated your changes

• Unit tests: ./pkg/clusteragent/admission/mutate/autoinstrumentation
• E2E TestSSISuite/TestRemoteConfig (HelmTargetWinsOverRemoteDeny, LastMatchingRemotePolicyWins)

Additional Notes

HelmTargetWinsOverRemoteDeny now expects the helm python workload to keep SSI when an RC deny matches targeted-namespace. That inverts #55057, where remote-before-local made the deny win.

[datadog-prod-us1-3]

🎯 Code Coverage (details)
• Patch Coverage: 89.36%
• Overall Coverage: 52.74% (-0.00%)

_{This comment will be updated automatically if new data arrives.
🔗 Commit SHA: a8fb51b | Docs | View more details | Give us feedback!}

[dd-octo-sts]

Files inventory check summary

File checks results against ancestor 240d736e:

Results for datadog-agent_7.84.0~devel.git.268.e2e3081.pipeline.132098738-1_amd64.deb:

No change detected

Results for datadog-iot-agent_7.84.0~devel.git.268.e2e3081.pipeline.132098738-1_amd64.deb:

No change detected

[cit-pr-commenter-54b7da]

Regression Detector

Regression Detector Results

Metrics dashboard
Target profiles
Job ID: 9e1c0cae-ccc3-4add-9ff6-d26a8485c668

Baseline: d725d90
Comparison: 5e98bef
Diff

Optimization Goals: ✅ No significant changes detected

Fine details of change detection per experiment

perf	experiment	goal	Δ mean %	Δ mean % CI	trials	links
➖	quality_gate_logs	% cpu utilization	+1.31	[+0.45, +2.17]	1	Logs bounds checks dashboard
➖	dsd_uds_10mb_3k_timestamped_contexts_cpu	% cpu utilization	+0.36	[+0.12, +0.61]	1	Logs
➖	quality_gate_metrics_logs	memory utilization	+0.18	[-0.05, +0.41]	1	Logs bounds checks dashboard
➖	quality_gate_security_mean_fs_load	memory utilization	+0.18	[+0.14, +0.21]	1	Logs bounds checks dashboard
➖	quality_gate_security_idle	memory utilization	+0.10	[+0.04, +0.15]	1	Logs bounds checks dashboard
➖	quality_gate_idle_all_features	memory utilization	-0.01	[-0.04, +0.02]	1	Logs bounds checks dashboard
➖	quality_gate_idle	memory utilization	-0.19	[-0.23, -0.14]	1	Logs bounds checks dashboard
➖	dsd_uds_10mb_3k_timestamped_contexts_memory	memory utilization	-0.35	[-0.56, -0.14]	1	Logs
➖	quality_gate_security_no_fs_load	memory utilization	-0.35	[-0.43, -0.27]	1	Logs bounds checks dashboard
➖	quality_gate_private_action_runner	memory utilization	-0.55	[-0.67, -0.43]	1	Logs bounds checks dashboard

Bounds Checks: ✅ Passed

perf	experiment	bounds_check_name	replicates_passed	observed_value	links
✅	quality_gate_idle	intake_connections	10/10	4 = 4	bounds checks dashboard
✅	quality_gate_idle	memory_usage	10/10	173.05MiB ≤ 179MiB	bounds checks dashboard
✅	quality_gate_idle	total_bytes_received	10/10	744.07KiB ≤ 819.20KiB	bounds checks dashboard
✅	quality_gate_idle_all_features	intake_connections	10/10	4 = 4	bounds checks dashboard
✅	quality_gate_idle_all_features	memory_usage	10/10	512.13MiB ≤ 537MiB	bounds checks dashboard
✅	quality_gate_idle_all_features	total_bytes_received	10/10	1.14MiB ≤ 1.25MiB	bounds checks dashboard
✅	quality_gate_logs	intake_connections	10/10	16 ≤ 40	bounds checks dashboard
✅	quality_gate_logs	memory_usage	10/10	210.28MiB ≤ 228MiB	bounds checks dashboard
✅	quality_gate_logs	missed_bytes	10/10	0B = 0B	bounds checks dashboard
✅	quality_gate_logs	total_bytes_received	10/10	263.46MiB ≤ 292MiB	bounds checks dashboard
✅	quality_gate_metrics_logs	cpu_usage	10/10	380.75 ≤ 2000	bounds checks dashboard
✅	quality_gate_metrics_logs	intake_connections	10/10	19 ≤ 40	bounds checks dashboard
✅	quality_gate_metrics_logs	memory_usage	10/10	408.08MiB ≤ 455MiB	bounds checks dashboard
✅	quality_gate_metrics_logs	missed_bytes	10/10	0B = 0B	bounds checks dashboard
✅	quality_gate_metrics_logs	total_bytes_received	10/10	0.94GiB ≤ 1.04GiB	bounds checks dashboard
✅	quality_gate_private_action_runner	memory_usage	10/10	72.22MiB ≤ 75MiB	bounds checks dashboard
✅	quality_gate_security_idle	cpu_usage	10/10	29.99 ≤ 100	bounds checks dashboard
✅	quality_gate_security_idle	memory_usage	10/10	334.98MiB ≤ 355MiB	bounds checks dashboard
✅	quality_gate_security_mean_fs_load	cpu_usage	10/10	63.74 ≤ 200	bounds checks dashboard
✅	quality_gate_security_mean_fs_load	memory_usage	10/10	304.24MiB ≤ 335MiB	bounds checks dashboard
✅	quality_gate_security_no_fs_load	cpu_usage	10/10	22.07 ≤ 100	bounds checks dashboard
✅	quality_gate_security_no_fs_load	memory_usage	10/10	313.98MiB ≤ 345MiB	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:

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".

CI Pass/Fail Decision

✅ Passed. All Quality Gates passed.

• quality_gate_metrics_logs, bounds check missed_bytes: 10/10 replicas passed. Gate passed.
• quality_gate_metrics_logs, bounds check cpu_usage: 10/10 replicas passed. Gate passed.
• quality_gate_metrics_logs, bounds check intake_connections: 10/10 replicas passed. Gate passed.
• quality_gate_metrics_logs, bounds check total_bytes_received: 10/10 replicas passed. Gate passed.
• quality_gate_metrics_logs, bounds check memory_usage: 10/10 replicas passed. Gate passed.
• quality_gate_logs, bounds check memory_usage: 10/10 replicas passed. Gate passed.
• quality_gate_logs, bounds check total_bytes_received: 10/10 replicas passed. Gate passed.
• quality_gate_logs, bounds check intake_connections: 10/10 replicas passed. Gate passed.
• quality_gate_logs, bounds check missed_bytes: 10/10 replicas passed. Gate passed.
• quality_gate_idle, bounds check memory_usage: 10/10 replicas passed. Gate passed.
• quality_gate_idle, bounds check intake_connections: 10/10 replicas passed. Gate passed.
• quality_gate_idle, bounds check total_bytes_received: 10/10 replicas passed. Gate passed.
• quality_gate_security_no_fs_load, bounds check memory_usage: 10/10 replicas passed. Gate passed.
• quality_gate_security_no_fs_load, bounds check cpu_usage: 10/10 replicas passed. Gate passed.
• quality_gate_idle_all_features, bounds check memory_usage: 10/10 replicas passed. Gate passed.
• quality_gate_idle_all_features, bounds check intake_connections: 10/10 replicas passed. Gate passed.
• quality_gate_idle_all_features, bounds check total_bytes_received: 10/10 replicas passed. Gate passed.
• quality_gate_security_mean_fs_load, bounds check memory_usage: 10/10 replicas passed. Gate passed.
• quality_gate_security_mean_fs_load, bounds check cpu_usage: 10/10 replicas passed. Gate passed.
• quality_gate_private_action_runner, bounds check memory_usage: 10/10 replicas passed. Gate passed.
• quality_gate_security_idle, bounds check cpu_usage: 10/10 replicas passed. Gate passed.
• quality_gate_security_idle, bounds check memory_usage: 10/10 replicas passed. Gate passed.

[dd-octo-sts]

Static quality checks

✅ Please find below the results from static quality gates
Comparison made with ancestor d08fa07
📊 Static Quality Gates Dashboard
🔗 SQG Job
SOME SIZE DELTAS ARE N/A (ANCESTOR METRICS NOT YET AVAILABLE). RETRY JOB

Successful checks

Info

	Quality gate	Change	Size (prev → curr → max)
✅	agent_deb_amd64	N/A	N/A → 761.671 → 764.840
✅	agent_deb_amd64_fips	N/A	N/A → 713.924 → 716.080
✅	agent_heroku_amd64	N/A	N/A → 312.967 → 319.080
✅	agent_msi	N/A	N/A → 648.347 → 661.050
✅	agent_rpm_amd64	N/A	N/A → 761.654 → 764.810
✅	agent_rpm_amd64_fips	N/A	N/A → 713.908 → 716.080
✅	agent_rpm_arm64	N/A	N/A → 737.552 → 738.310
✅	agent_rpm_arm64_fips	N/A	N/A → 693.143 → 694.350
✅	agent_suse_amd64	N/A	N/A → 761.654 → 764.810
✅	agent_suse_amd64_fips	N/A	N/A → 713.908 → 716.080
✅	agent_suse_arm64	N/A	N/A → 737.552 → 738.310
✅	agent_suse_arm64_fips	N/A	N/A → 693.143 → 694.350
✅	docker_agent_amd64	N/A	N/A → 820.687 → 821.000
✅	docker_agent_arm64	N/A	N/A → 821.620 → 821.980
✅	docker_agent_jmx_amd64	N/A	N/A → 1011.599 → 1011.910
✅	docker_agent_jmx_arm64	N/A	N/A → 1001.312 → 1001.680
✅	docker_cluster_agent_amd64	N/A	N/A → 211.261 → 212.130
✅	docker_cluster_agent_arm64	N/A	N/A → 224.398 → 225.220
✅	docker_cws_instrumentation_amd64	N/A	N/A → 7.439 → 8.400
✅	docker_cws_instrumentation_arm64	N/A	N/A → 6.877 → 7.110
✅	docker_dogstatsd_amd64	N/A	N/A → 39.545 → 40.440
✅	docker_dogstatsd_arm64	N/A	N/A → 37.623 → 38.580
✅	docker_host_profiler_amd64	N/A	N/A → 306.862 → 317.700
✅	docker_host_profiler_arm64	N/A	N/A → 318.167 → 328.970
✅	dogstatsd_deb_amd64	N/A	N/A → 30.290 → 31.210
✅	dogstatsd_deb_arm64	N/A	N/A → 28.302 → 29.590
✅	dogstatsd_rpm_amd64	N/A	N/A → 30.290 → 31.210
✅	dogstatsd_suse_amd64	N/A	N/A → 30.290 → 31.210
✅	iot_agent_deb_amd64	N/A	N/A → 46.568 → 47.550
✅	iot_agent_deb_arm64	N/A	N/A → 43.204 → 44.220
✅	iot_agent_deb_armhf	N/A	N/A → 44.015 → 45.020
✅	iot_agent_rpm_amd64	N/A	N/A → 46.568 → 47.550
✅	iot_agent_suse_amd64	N/A	N/A → 46.567 → 47.550

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 9e337ef33d

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[maattdd]

on | present | policies | first (as in the yaml) matching target, else last matching policy, else nothing

Do we want static target to win over RC policies ?

[iamluc]

Do we want static target to win over RC policies ?

If we concatenate RC and static targets into one last-wins list, static targets have to come after the RC rules. Otherwise (static first), the RC catch-all always overrides them, since the last matching rule wins.

Also, static targets can only allow instrumentation—they can never block it. This means:

RC = Allow + exclusions: a matching static target can re-allow an excluded workload (static targets win).
RC = Block + exceptions: a matching static target will always win.

I think evaluating the policies twice—first the static targets, then the RC rules—results in a much simpler mental model. That's a very important point, IMHO.

[iamluc]

@codex review

[chatgpt-codex-connector]

Codex Review: Didn't find any major issues. 🎉

Reviewed commit: e2e308115d

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[maattdd]

We should discuss what is the most useful behavior for the case of target allow + explicit RC deny.

Otherwise, LGTM