Skip to content

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

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
gh-worker-dd-mergequeue-cf854d merged 3 commits into from
Aug 5, 2026
Merged

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

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
20ea0f4
fix(gpu): gate the event monitor on enable_ebpf_probes
mbertrone Aug 3, 2026
b75f160
rm release note
mbertrone Aug 4, 2026
a8d1c7d
Merge branch 'main' into matteo.bertrone/gpu-gate-event-monitor-on-eb…
mbertrone Aug 5, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 5 additions & 1 deletion cmd/system-probe/modules/eventmonitor.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -133,7 +133,11 @@ func createEventMonitorModule(_ *sysconfigtypes.Config, deps module.FactoryDepen
}

gpucfg := gpuconfig.New()
if gpucfg.Enabled {
// Only the eBPF probes consume these events, so skip the consumer entirely
// when they are disabled. Kept in sync with the module gate in
// pkg/system-probe/config, which does not enable the event monitor for GPU
// monitoring unless the probes are enabled too.
if gpucfg.Enabled && gpucfg.EnableEBPFProbes {
err := createGPUProcessEventConsumer(evm)
if err != nil {
return nil, fmt.Errorf("cannot create event consumer for GPU: %w", err)
Expand Down
5 changes: 4 additions & 1 deletion pkg/system-probe/config/config.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -127,6 +127,9 @@ func load() (*types.Config, error) {
eudmEnabled := coreCfg.GetString("infrastructure_mode") == "end_user_device"
csmEnabled := cfg.GetBool(secNS("enabled"))
gpuEnabled := cfg.GetBool(gpuNS("enabled"))
// The GPU module only consumes process events when its eBPF probes are
// loaded, so the event monitor is only needed for that combination.
gpuEBPFProbesEnabled := gpuEnabled && cfg.GetBool(gpuNS("enable_ebpf_probes"))
diEnabled := cfg.GetBool(diNS("enabled"))
swEnabled := coreCfg.GetBool(swNS("enabled"))
discoveryServiceMapEnabled := cfg.GetBool(discoveryNS("service_map", "enabled"))
Expand All @@ -145,7 +148,7 @@ func load() (*types.Config, error) {
coreCfg.GetBool("sbom.enrichment.usage.enabled") ||
(usmEnabled && cfg.GetBool(smNS("enable_event_stream"))) ||
(c.ModuleIsEnabled(NetworkTracerModule) && cfg.GetBool(evNS("network_process.enabled"))) ||
gpuEnabled ||
gpuEBPFProbesEnabled ||
diEnabled {
c.EnabledModules[EventMonitorModule] = struct{}{}
}
Expand Down
10 changes: 9 additions & 1 deletion pkg/system-probe/config/config_test.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -23,6 +23,7 @@ func TestEventMonitor(t *testing.T) {

for i, tc := range []struct {
cws, fim, networkEvents, gpu bool
gpuEBPFProbes bool
usmEvents bool
enabled bool
}{
Expand All @@ -34,7 +35,11 @@ func TestEventMonitor(t *testing.T) {
{cws: false, fim: true, networkEvents: true, enabled: true},
{cws: true, fim: false, networkEvents: true, enabled: true},
{cws: true, fim: true, networkEvents: true, enabled: true},
{cws: false, fim: false, networkEvents: false, gpu: true, enabled: true},
// GPU monitoring only needs the event monitor to feed its eBPF probes,
// so both settings have to be enabled for the module to be pulled in.
{cws: false, fim: false, networkEvents: false, gpu: true, gpuEBPFProbes: true, enabled: true},
{cws: false, fim: false, networkEvents: false, gpu: true, gpuEBPFProbes: false, enabled: false},
{cws: false, fim: false, networkEvents: false, gpu: false, gpuEBPFProbes: true, enabled: false},
{usmEvents: true, enabled: true},
} {
t.Run(strconv.Itoa(i), func(t *testing.T) {
Expand All @@ -44,6 +49,9 @@ func TestEventMonitor(t *testing.T) {
t.Setenv("DD_SYSTEM_PROBE_EVENT_MONITORING_NETWORK_PROCESS_ENABLED", strconv.FormatBool(tc.networkEvents))
t.Setenv("DD_SYSTEM_PROBE_NETWORK_ENABLED", strconv.FormatBool(tc.networkEvents))
t.Setenv("DD_GPU_MONITORING_ENABLED", strconv.FormatBool(tc.gpu))
// Set explicitly rather than relying on the default, which is
// subject to change as the eBPF probes are deprecated.
t.Setenv("DD_GPU_MONITORING_ENABLE_EBPF_PROBES", strconv.FormatBool(tc.gpuEBPFProbes))
t.Setenv("DD_SYSTEM_PROBE_SERVICE_MONITORING_ENABLED", strconv.FormatBool(tc.usmEvents))
t.Setenv("DD_SERVICE_MONITORING_CONFIG_ENABLE_EVENT_STREAM", strconv.FormatBool(tc.usmEvents))

Expand Down
Loading