Skip to content

fix(monitor): use IsValidUUID to skip uninitialised device UUIDs in scrape path #2465

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
hami-robot merged 1 commit into from
Aug 11, 2026
Merged

fix(monitor): use IsValidUUID to skip uninitialised device UUIDs in scrape path #2465

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
20 changes: 11 additions & 9 deletions cmd/vGPUmonitor/feedback_test.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -64,15 +64,17 @@ func (s *stubInfo) DeviceMemoryOffset(int) uint64 { return 0 }
func (s *stubInfo) DeviceMemoryTotal(i int) uint64 { return slot(s.total, i) }
func (s *stubInfo) DeviceSmUtil(i int) uint64 { return slot(s.smUtil, i) }
func (s *stubInfo) SetDeviceSmLimit(uint64) {}
func (s *stubInfo) IsValidUUID(i int) bool { return i < len(s.uuids) }
func (s *stubInfo) DeviceMemoryLimit(i int) uint64 { return slot(s.limit, i) }
func (s *stubInfo) SetDeviceMemoryLimit(uint64) {}
func (s *stubInfo) LastKernelTime() int64 { return s.lastKernel }
func (s *stubInfo) GetPriority() int { return s.priority }
func (s *stubInfo) GetRecentKernel() int32 { return 1 }
func (s *stubInfo) SetRecentKernel(int32) {}
func (s *stubInfo) GetUtilizationSwitch() int32 { return 0 }
func (s *stubInfo) SetUtilizationSwitch(int32) {}
func (s *stubInfo) IsValidUUID(i int) bool {
return i >= 0 && i < len(s.uuids) && len(s.uuids[i]) > 0 && s.uuids[i][0] != 0
}
Comment thread
coderabbitai[bot] marked this conversation as resolved.
func (s *stubInfo) DeviceMemoryLimit(i int) uint64 { return slot(s.limit, i) }
func (s *stubInfo) SetDeviceMemoryLimit(uint64) {}
func (s *stubInfo) LastKernelTime() int64 { return s.lastKernel }
func (s *stubInfo) GetPriority() int { return s.priority }
func (s *stubInfo) GetRecentKernel() int32 { return 1 }
func (s *stubInfo) SetRecentKernel(int32) {}
func (s *stubInfo) GetUtilizationSwitch() int32 { return 0 }
func (s *stubInfo) SetUtilizationSwitch(int32) {}

func TestCheckFunctionsHighPriority(t *testing.T) {
sw := map[string]UtilizationPerDevice{"gpu-0": {0, 1}}
Expand Down
9 changes: 4 additions & 5 deletions cmd/vGPUmonitor/metrics.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -428,14 +428,13 @@ func (cc ClusterManagerCollector) collectContainerMetrics(ch chan<- prometheus.M

// Iterate through each device
for i := range c.Info.DeviceNum() {
uuid := c.Info.DeviceUUID(i)
if len(uuid) < 40 {
klog.Warningf("Device %d in Pod %s/%s, Container %s has invalid UUID length %d (shared memory not yet initialised); skipping until next scrape", i, pod.Namespace, pod.Name, ctr.Name, len(uuid))
if !c.Info.IsValidUUID(i) {
klog.Warningf("Device %d in Pod %s/%s, Container %s UUID not yet initialised; skipping until next scrape", i, pod.Namespace, pod.Name, ctr.Name)
continue
}
uuid = uuid[0:40] // Ensure UUID is truncated to 40 characters
uuid := c.Info.DeviceUUID(i)[0:40]
if !utf8.ValidString(uuid) {
klog.Warningf("Device %d in Pod %s/%s, Container %s has invalid UTF-8 UUID (shared memory not yet initialised); skipping until next scrape", i, pod.Namespace, pod.Name, ctr.Name)
klog.Warningf("Device %d in Pod %s/%s, Container %s has invalid UTF-8 UUID; skipping until next scrape", i, pod.Namespace, pod.Name, ctr.Name)
continue
}

Expand Down
8 changes: 4 additions & 4 deletions cmd/vGPUmonitor/metrics_container_test.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -133,13 +133,13 @@ func TestCollectContainerMetricsBadInput(t *testing.T) {
if _, err := collectContainer(t, &nvidia.ContainerUsage{}, 0); err == nil {
t.Error("nil ContainerUsage.Info should return an error")
}
short := &nvidia.ContainerUsage{Info: &stubInfo{uuids: []string{"gpu-short"}}}
metrics, err := collectContainer(t, short, 0)
uninit := &nvidia.ContainerUsage{Info: &stubInfo{uuids: []string{"\x00" + strings.Repeat("\x00", 39)}}}
metrics, err := collectContainer(t, uninit, 0)
if err != nil {
t.Fatalf("a UUID shorter than 40 chars should be skipped, not error: %v", err)
t.Fatalf("uninitialized UUID should be skipped, not error: %v", err)
}
if len(metrics) != 0 {
t.Errorf("got %d metrics for a short UUID, want 0", len(metrics))
t.Errorf("got %d metrics for an uninitialized UUID, want 0", len(metrics))
}
}

Expand Down
Loading