fix(gpu): gate the event monitor on enable_ebpf_probes by mbertrone · Pull Request #54372 · DataDog/datadog-agent

mbertrone · 2026-08-03T15:36:43Z

What does this PR do?

Gates the system-probe event monitor on gpu_monitoring.enable_ebpf_probes in
addition to gpu_monitoring.enabled, in both the module-enablement logic
(pkg/system-probe/config) and the GPU process-consumer creation
(cmd/system-probe/modules/eventmonitor.go).

The GPU monitoring module itself remains gated on gpu_monitoring.enabled, so
NVML-based collection is unchanged.

Motivation

The event monitor exists only to feed process exec/exit events to the GPU eBPF
probes. Keyed on gpu_monitoring.enabled alone, a host running GPU monitoring
through NVML only still loaded the event monitor's eBPF programs and built a
process consumer that nothing read.

Follow-up to the discussion on #54291.

Describe how you validated your changes

Extended TestEventMonitor with the enabled × enable_ebpf_probes matrix,
setting both explicitly rather than relying on the default (which is changing as
the probes are deprecated). Verified that with the probes disabled the gpu
module is still enabled while event_monitor is not:

`gpu_monitoring.enabled`	`enable_ebpf_probes`	enabled modules
true	true	`discovery, event_monitor, gpu`
true	false	`discovery, gpu`
false	true	`discovery`

Also built cmd/system-probe/modules with linux_bpf,nvml to cover the
configuration that includes both eventmonitor.go and the real gpu.go.

Additional Notes

Both gates had to move together to stay consistent with the
EnableEBPFProbes && processEventConsumer == nil guard in the GPU module
factory.

Note that gpu_monitoring.enable_ebpf_probes still defaults to true on main,
so this is a no-op for default configurations until #54291 lands; it currently
only affects deployments that set it to false explicitly.

🤖 Generated with Claude Code

datadog-prod-us1-3 · 2026-08-03T16:02:54Z

🎯 Code Coverage (details)
• Patch Coverage: 66.67%
• Overall Coverage: 52.23% (+1.57%)

_{This comment will be updated automatically if new data arrives.

🔗 Commit SHA: a8d1c7d | Docs | Datadog PR Page | Give us feedback!}

dd-octo-sts · 2026-08-03T16:24:18Z

Files inventory check summary

File checks results against ancestor daf42da4:

Results for datadog-agent_7.83.0~devel.git.577.a8d1c7d.pipeline.129069201-1_amd64.deb:

No change detected

dd-octo-sts · 2026-08-03T16:33:53Z

Static quality checks

✅ Please find below the results from static quality gates
Comparison made with ancestor daf42da
📊 Static Quality Gates Dashboard
🔗 SQG Job

Successful checks

Info

	Quality gate	Change	Size (prev → curr → max)
✅	agent_deb_amd64	+8.0 KiB (0.00% increase, -0.15% of buffer)	757.651 → 757.659 → 762.870
✅	agent_msi	+2.5 KiB (0.00% increase, -0.01% of buffer)	639.345 → 639.347 → 657.290
✅	agent_rpm_amd64	+8.0 KiB (0.00% increase, -0.15% of buffer)	757.635 → 757.642 → 762.840
✅	agent_rpm_arm64_fips	+4.0 KiB (0.00% increase, -0.25% of buffer)	690.659 → 690.663 → 692.250
✅	agent_suse_amd64	+8.0 KiB (0.00% increase, -0.15% of buffer)	757.635 → 757.642 → 762.840
✅	agent_suse_arm64_fips	+4.0 KiB (0.00% increase, -0.25% of buffer)	690.659 → 690.663 → 692.250
✅	docker_agent_amd64	+8.0 KiB (0.00% increase, -0.36% of buffer)	816.305 → 816.313 → 818.460
✅	docker_agent_jmx_amd64	+8.0 KiB (0.00% increase, -0.39% of buffer)	1007.203 → 1007.211 → 1009.220

25 successful checks with minimal change (< 2 KiB)

	Quality gate	Current Size
✅	agent_deb_amd64_fips	711.296 MiB
✅	agent_heroku_amd64	311.886 MiB
✅	agent_rpm_amd64_fips	711.279 MiB
✅	agent_rpm_arm64	733.739 MiB
✅	agent_suse_amd64_fips	711.279 MiB
✅	agent_suse_arm64	733.739 MiB
✅	docker_agent_arm64	817.473 MiB
✅	docker_agent_jmx_arm64	997.023 MiB
✅	docker_cluster_agent_amd64	210.030 MiB
✅	docker_cluster_agent_arm64	223.077 MiB
✅	docker_cws_instrumentation_amd64	7.439 MiB
✅	docker_cws_instrumentation_arm64	6.877 MiB
✅	docker_dogstatsd_amd64	39.312 MiB
✅	docker_dogstatsd_arm64	37.433 MiB
✅	docker_host_profiler_amd64	305.307 MiB
✅	docker_host_profiler_arm64	316.620 MiB
✅	dogstatsd_deb_amd64	30.053 MiB
✅	dogstatsd_deb_arm64	28.089 MiB
✅	dogstatsd_rpm_amd64	30.053 MiB
✅	dogstatsd_suse_amd64	30.053 MiB
✅	iot_agent_deb_amd64	46.290 MiB
✅	iot_agent_deb_arm64	42.973 MiB
✅	iot_agent_deb_armhf	43.738 MiB
✅	iot_agent_rpm_amd64	46.290 MiB
✅	iot_agent_suse_amd64	46.289 MiB

cit-pr-commenter-54b7da · 2026-08-03T16:53:47Z

Regression Detector

Regression Detector Results

Metrics dashboard
Target profiles
Run ID: 6319b02f-dfa5-43e7-b60e-ff5133347566

Baseline: 822ef92
Comparison: 8657403
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_metrics_logs	memory utilization	+1.66	[+1.41, +1.91]	1	Logs bounds checks dashboard
➖	quality_gate_security_no_fs_load	memory utilization	+0.33	[+0.23, +0.43]	1	Logs bounds checks dashboard
➖	quality_gate_private_action_runner	memory utilization	+0.25	[+0.13, +0.37]	1	Logs bounds checks dashboard
➖	quality_gate_idle	memory utilization	+0.11	[+0.07, +0.16]	1	Logs bounds checks dashboard
➖	quality_gate_logs	% cpu utilization	+0.04	[-0.80, +0.88]	1	Logs bounds checks dashboard
➖	quality_gate_security_idle	memory utilization	+0.02	[-0.04, +0.07]	1	Logs bounds checks dashboard
➖	quality_gate_idle_all_features	memory utilization	-0.30	[-0.34, -0.26]	1	Logs bounds checks dashboard
➖	quality_gate_security_mean_fs_load	memory utilization	-0.34	[-0.37, -0.30]	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	3 ≤ 4	bounds checks dashboard
✅	quality_gate_idle	memory_usage	10/10	150.16MiB ≤ 154MiB	bounds checks dashboard
✅	quality_gate_idle	total_bytes_received	10/10	734.52KiB ≤ 819.20KiB	bounds checks dashboard
✅	quality_gate_idle_all_features	intake_connections	10/10	3 ≤ 4	bounds checks dashboard
✅	quality_gate_idle_all_features	memory_usage	10/10	494.11MiB ≤ 512MiB	bounds checks dashboard
✅	quality_gate_idle_all_features	total_bytes_received	10/10	1.12MiB ≤ 1.25MiB	bounds checks dashboard
✅	quality_gate_logs	intake_connections	10/10	4 ≤ 6	bounds checks dashboard
✅	quality_gate_logs	memory_usage	10/10	186.44MiB ≤ 195MiB	bounds checks dashboard
✅	quality_gate_logs	missed_bytes	10/10	0B = 0B	bounds checks dashboard
✅	quality_gate_logs	total_bytes_received	10/10	264.15MiB ≤ 292MiB	bounds checks dashboard
✅	quality_gate_metrics_logs	cpu_usage	10/10	401.75 ≤ 2000	bounds checks dashboard
✅	quality_gate_metrics_logs	intake_connections	10/10	3 ≤ 6	bounds checks dashboard
✅	quality_gate_metrics_logs	memory_usage	10/10	410.28MiB ≤ 430MiB	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	72.55MiB ≤ 75MiB	bounds checks dashboard
✅	quality_gate_security_idle	cpu_usage	10/10	26.67 ≤ 100	bounds checks dashboard
✅	quality_gate_security_idle	memory_usage	10/10	303.81MiB ≤ 330MiB	bounds checks dashboard
✅	quality_gate_security_mean_fs_load	cpu_usage	10/10	60.45 ≤ 200	bounds checks dashboard
✅	quality_gate_security_mean_fs_load	memory_usage	10/10	278.15MiB ≤ 310MiB	bounds checks dashboard
✅	quality_gate_security_no_fs_load	cpu_usage	10/10	20.57 ≤ 100	bounds checks dashboard
✅	quality_gate_security_no_fs_load	memory_usage	10/10	288.03MiB ≤ 320MiB	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".

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 Profiling Replicate Execution Failures (ddprof)

Note: Profiling replicas may still be executing. See the debug dashboard for up to date status.

Experiment	Variant	Replicates	Failure	Debug Dashboard
quality_gate_idle_all_features	baseline	10	Oom killed	Debug Dashboard
quality_gate_idle_all_features	comparison	10	Oom killed	Debug Dashboard
quality_gate_logs	baseline	10	Oom killed	Debug Dashboard
quality_gate_logs	comparison	10	Oom killed	Debug Dashboard
quality_gate_metrics_logs	baseline	10	Oom killed	Debug Dashboard
quality_gate_metrics_logs	comparison	10	Oom killed	Debug Dashboard
quality_gate_security_mean_fs_load	baseline	10	Crashed (exit code: 134)	Debug Dashboard
quality_gate_security_mean_fs_load	comparison	10	Crashed (exit code: 134)	Debug Dashboard
quality_gate_security_no_fs_load	baseline	10	Oom killed	Debug Dashboard

CI Pass/Fail Decision

✅ Passed. All Quality Gates passed.

quality_gate_security_no_fs_load, bounds check cpu_usage: 10/10 replicas passed. Gate passed.
quality_gate_security_no_fs_load, bounds check memory_usage: 10/10 replicas passed. Gate passed.
quality_gate_logs, bounds check missed_bytes: 10/10 replicas passed. Gate passed.
quality_gate_logs, bounds check intake_connections: 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_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_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_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_metrics_logs, bounds check total_bytes_received: 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 missed_bytes: 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 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.

The event monitor is only used to feed the GPU monitoring eBPF probes with process exec/exit events, but both the system-probe module gate and the consumer creation were keyed on gpu_monitoring.enabled alone. A host running GPU monitoring through NVML only therefore still loaded the event monitor's eBPF programs and built a process consumer that nothing read. Gate both on gpu_monitoring.enable_ebpf_probes as well. The two have to move together to stay consistent with the EnableEBPFProbes && processEventConsumer == nil guard in the GPU module factory. The GPU module itself is still gated on gpu_monitoring.enabled, so NVML-based collection is unaffected. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

not needed

…pf-probes

mbertrone · 2026-08-05T07:36:48Z

/merge

gh-worker-devflow-routing-ef8351 · 2026-08-05T07:36:53Z

View all feedbacks in Devflow UI.

2026-08-05 07:36:53 UTC ℹ️ Start processing command /merge

2026-08-05 07:36:59 UTC ℹ️ MergeQueue: waiting for PR to be ready

This pull request is not mergeable according to GitHub. Common reasons include pending required checks, missing approvals, or merge conflicts — but it could also be blocked by other repository rules or settings.
It will be added to the queue as soon as checks pass and/or get approvals. View in MergeQueue UI.
Note: if you pushed new commits since the last approval, you may need additional approval.
You can remove it from the waiting list with /remove command.

2026-08-05 09:10:06 UTC ℹ️ MergeQueue: merge request added to the queue

The expected merge time in main is approximately 2h (p90).

2026-08-05 09:57:21 UTC ℹ️ MergeQueue: This merge request was merged

### What does this PR do? Gates the system-probe event monitor on `gpu_monitoring.enable_ebpf_probes` in addition to `gpu_monitoring.enabled`, in both the module-enablement logic (`pkg/system-probe/config`) and the GPU process-consumer creation (`cmd/system-probe/modules/eventmonitor.go`). The GPU monitoring module itself remains gated on `gpu_monitoring.enabled`, so NVML-based collection is unchanged. ### Motivation The event monitor exists only to feed process exec/exit events to the GPU eBPF probes. Keyed on `gpu_monitoring.enabled` alone, a host running GPU monitoring through NVML only still loaded the event monitor's eBPF programs and built a process consumer that nothing read. Follow-up to the discussion on #54291. ### Describe how you validated your changes Extended `TestEventMonitor` with the `enabled` × `enable_ebpf_probes` matrix, setting both explicitly rather than relying on the default (which is changing as the probes are deprecated). Verified that with the probes disabled the `gpu` module is still enabled while `event_monitor` is not: | `gpu_monitoring.enabled` | `enable_ebpf_probes` | enabled modules | | --- | --- | --- | | true | true | `discovery, event_monitor, gpu` | | true | false | `discovery, gpu` | | false | true | `discovery` | Also built `cmd/system-probe/modules` with `linux_bpf,nvml` to cover the configuration that includes both `eventmonitor.go` and the real `gpu.go`. ### Additional Notes Both gates had to move together to stay consistent with the `EnableEBPFProbes && processEventConsumer == nil` guard in the GPU module factory. Note that `gpu_monitoring.enable_ebpf_probes` still defaults to `true` on `main`, so this is a no-op for default configurations until #54291 lands; it currently only affects deployments that set it to `false` explicitly. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-authored-by: matteo.bertrone <matteo.bertrone@datadoghq.com> 8657403

github-actions Bot added the component/system-probe label Aug 3, 2026

dd-octo-sts Bot added internal Identify a non-fork PR team/ebpf-platform team/agent-security labels Aug 3, 2026

github-actions Bot added the medium review PR review might take time label Aug 3, 2026

mbertrone marked this pull request as ready for review August 3, 2026 15:38

mbertrone requested review from a team as code owners August 3, 2026 15:38

gjulianm approved these changes Aug 3, 2026

View reviewed changes

Comment thread releasenotes/notes/gpu-event-monitor-gated-on-ebpf-probes-8935138775ed9a76.yaml Outdated

spikat approved these changes Aug 4, 2026

View reviewed changes

mbertrone self-assigned this Aug 4, 2026

mbertrone added the qa/skip-qa label Aug 4, 2026

github-actions Bot added short review PR is simple enough to be reviewed quickly and removed medium review PR review might take time labels Aug 4, 2026

mbertrone added qa/done QA done before merge and regressions are covered by tests changelog/no-changelog No changelog entry needed and removed qa/skip-qa labels Aug 4, 2026

mbertrone and others added 2 commits August 4, 2026 18:03

rm release note

b75f160

not needed

mbertrone force-pushed the matteo.bertrone/gpu-gate-event-monitor-on-ebpf-probes branch from 1d00abb to b75f160 Compare August 4, 2026 16:47

Merge branch 'main' into matteo.bertrone/gpu-gate-event-monitor-on-eb…

a8d1c7d

…pf-probes

gh-worker-dd-mergequeue-cf854d Bot merged commit 8657403 into main Aug 5, 2026
398 checks passed

gh-worker-dd-mergequeue-cf854d Bot deleted the matteo.bertrone/gpu-gate-event-monitor-on-ebpf-probes branch August 5, 2026 09:57

github-actions Bot added this to the 7.83.0 milestone Aug 5, 2026

Navigation Menu

Uh oh!

fix(gpu): gate the event monitor on enable_ebpf_probes - #54372

fix(gpu): gate the event monitor on enable_ebpf_probes#54372
gh-worker-dd-mergequeue-cf854d[bot] merged 3 commits into
mainfrom
matteo.bertrone/gpu-gate-event-monitor-on-ebpf-probes

mbertrone commented Aug 3, 2026

Uh oh!

Uh oh!

datadog-prod-us1-3 Bot commented Aug 3, 2026 •

edited

Loading

Uh oh!

dd-octo-sts Bot commented Aug 3, 2026 •

edited

Loading

Uh oh!

dd-octo-sts Bot commented Aug 3, 2026 •

edited

Loading

Info

Uh oh!

cit-pr-commenter-54b7da Bot commented Aug 3, 2026 •

edited

Loading

Fine details of change detection per experiment

Bounds Checks: ✅ Passed

Explanation

Replicate Execution Details

❌ Retried Profiling Replicate Execution Failures (ddprof)

Uh oh!

mbertrone commented Aug 5, 2026

Uh oh!

gh-worker-devflow-routing-ef8351 Bot commented Aug 5, 2026 •

edited

Loading

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

Uh oh!

Conversation

mbertrone commented Aug 3, 2026

What does this PR do?

Motivation

Describe how you validated your changes

Additional Notes

Uh oh!

Uh oh!

datadog-prod-us1-3 Bot commented Aug 3, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

dd-octo-sts Bot commented Aug 3, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Files inventory check summary

Results for datadog-agent_7.83.0~devel.git.577.a8d1c7d.pipeline.129069201-1_amd64.deb:

Uh oh!

dd-octo-sts Bot commented Aug 3, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Static quality checks

Info

Uh oh!

cit-pr-commenter-54b7da Bot commented Aug 3, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Regression Detector

Regression Detector Results

Optimization Goals: ✅ No significant changes detected

Fine details of change detection per experiment

Bounds Checks: ✅ Passed

Explanation

Replicate Execution Details

❌ Retried Profiling Replicate Execution Failures (ddprof)

CI Pass/Fail Decision

Uh oh!

mbertrone commented Aug 5, 2026

Uh oh!

gh-worker-devflow-routing-ef8351 Bot commented Aug 5, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

datadog-prod-us1-3 Bot commented Aug 3, 2026 •

edited

Loading

dd-octo-sts Bot commented Aug 3, 2026 •

edited

Loading

dd-octo-sts Bot commented Aug 3, 2026 •

edited

Loading

cit-pr-commenter-54b7da Bot commented Aug 3, 2026 •

edited

Loading

gh-worker-devflow-routing-ef8351 Bot commented Aug 5, 2026 •

edited

Loading