pkg/dcgmexporter/gpu_collector.go: include a err_msg label in metric DCGM_FI_DEV_XID_ERRORS

pkg/dcgmexporter/gpu_collector.go

@@ -20,6 +20,7 @@ import (
 	"errors"
 	"fmt"
 	"os"
+	"strconv"
 	"strings"
 
 	"github.com/NVIDIA/go-dcgm/pkg/dcgm"
@@ -31,8 +32,8 @@ type DCGMCollectorConstructor func([]Counter, string, *Config, FieldEntityGroupT
 func NewDCGMCollector(c []Counter,
 	hostname string,
 	config *Config,
-	fieldEntityGroupTypeSystemInfo FieldEntityGroupTypeSystemInfoItem) (*DCGMCollector, func(), error) {
-
+	fieldEntityGroupTypeSystemInfo FieldEntityGroupTypeSystemInfoItem,
+) (*DCGMCollector, func(), error) {
 	if fieldEntityGroupTypeSystemInfo.isEmpty() {
 		return nil, func() {}, errors.New("fieldEntityGroupTypeSystemInfo is empty")
 	}
@@ -163,7 +164,8 @@ func FindCounterField(c []Counter, fieldId uint) (Counter, error) {
 }
 
 func ToSwitchMetric(metrics MetricsByCounter,
-	values []dcgm.FieldValue_v1, c []Counter, mi MonitoringInfo, useOld bool, hostname string) {
+	values []dcgm.FieldValue_v1, c []Counter, mi MonitoringInfo, useOld bool, hostname string,
+) {
 	labels := map[string]string{}
 
 	for _, val := range values {
@@ -206,8 +208,9 @@ func ToSwitchMetric(metrics MetricsByCounter,
 }
 
 func ToCPUMetric(metrics MetricsByCounter,
-	values []dcgm.FieldValue_v1, c []Counter, mi MonitoringInfo, useOld bool, hostname string) {
-	var labels = map[string]string{}
+	values []dcgm.FieldValue_v1, c []Counter, mi MonitoringInfo, useOld bool, hostname string,
+) {
+	labels := map[string]string{}
 
 	for _, val := range values {
 		v := ToString(val)
@@ -258,7 +261,7 @@ func ToMetric(
 	hostname string,
 	replaceBlanksInModelName bool,
 ) {
-	var labels = map[string]string{}
+	labels := map[string]string{}
 
 	for _, val := range values {
 		v := ToString(val)
@@ -283,6 +286,17 @@ func ToMetric(
 
 		gpuModel := getGPUModel(d, replaceBlanksInModelName)
 
+		attrs := map[string]string{}
+		if counter.FieldID == dcgm.DCGM_FI_DEV_XID_ERRORS {
+			errCode := int(val.Int64())
+			attrs["err_code"] = strconv.Itoa(errCode)
+			if 0 < errCode && errCode < len(xidErrCodeToText) {
+				attrs["err_msg"] = xidErrCodeToText[errCode]
+			} else {
+				attrs["err_msg"] = "Unknown Error"
+			}
+		}
+
 		m := Metric{
 			Counter: counter,
 			Value:   v,
@@ -295,7 +309,7 @@ func ToMetric(
 			Hostname:     hostname,
 
 			Labels:     labels,
-			Attributes: map[string]string{},
+			Attributes: attrs,
 		}
 		if instanceInfo != nil {
 			m.MigProfile = instanceInfo.ProfileName

pkg/dcgmexporter/gpu_collector_test.go

@@ -257,7 +257,6 @@ func testDCGMCPUCollector(t *testing.T, counters []Counter) (*DCGMCollector, fun
 }
 
 func TestToMetric(t *testing.T) {
-
 	fieldValue := [4096]byte{}
 	fieldValue[0] = 42
 	values := []dcgm.FieldValue_v1{
@@ -315,6 +314,70 @@ func TestToMetric(t *testing.T) {
 	}
 }
 
+func TestToMetricWhenDCGM_FI_DEV_XID_ERRORSField(t *testing.T) {
+	c := []Counter{
+		{
+			FieldID:   dcgm.DCGM_FI_DEV_XID_ERRORS,
+			FieldName: "DCGM_FI_DEV_GPU_TEMP",
+			PromType:  "gauge",
+			Help:      "Temperature Help info",
+		},
+	}
+
+	d := dcgm.Device{
+		UUID: "fake0",
+		Identifiers: dcgm.DeviceIdentifiers{
+			Model: "NVIDIA T400 4GB",
+		},
+	}
+
+	var instanceInfo *GPUInstanceInfo = nil
+
+	type testCase struct {
+		name        string
+		fieldValue  byte
+		expectedErr string
+	}
+
+	testCases := []testCase{
+		{
+			name:        "when DCGM_FI_DEV_XID_ERRORS has known value",
+			fieldValue:  42,
+			expectedErr: "Video processor exception",
+		},
+		{
+			name:        "when DCGM_FI_DEV_XID_ERRORS has unknown value",
+			fieldValue:  255,
+			expectedErr: "Unknown Error",
+		},
+	}
+
+	for _, tc := range testCases {
+		t.Run(tc.name, func(t *testing.T) {
+			fieldValue := [4096]byte{}
+			fieldValue[0] = tc.fieldValue
+			values := []dcgm.FieldValue_v1{
+				{
+					FieldId:   dcgm.DCGM_FI_DEV_XID_ERRORS,
+					FieldType: dcgm.DCGM_FT_INT64,
+					Value:     fieldValue,
+				},
+			}
+
+			metrics := make(map[Counter][]Metric)
+			ToMetric(metrics, values, c, d, instanceInfo, false, "", false)
+			assert.Len(t, metrics, 1)
+			// We get metric value with 0 index
+			metricValues := metrics[reflect.ValueOf(metrics).MapKeys()[0].Interface().(Counter)]
+			assert.Equal(t, fmt.Sprint(tc.fieldValue), metricValues[0].Value)
+			assert.Contains(t, metricValues[0].Attributes, "err_code")
+			assert.Equal(t, fmt.Sprint(tc.fieldValue), metricValues[0].Attributes["err_code"])
+			assert.Contains(t, metricValues[0].Attributes, "err_msg")
+			assert.Equal(t, tc.expectedErr, metricValues[0].Attributes["err_msg"])
+		})
+	}
+}
+
 func TestGPUCollector_GetMetrics(t *testing.T) {
 	teardownTest := setupTest(t)
 	defer teardownTest(t)

pkg/dcgmexporter/xid_errors.go

@@ -0,0 +1,165 @@
+/*
+ * Copyright (c) 2024, NVIDIA CORPORATION.  All rights reserved.
+ *
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+package dcgmexporter
+
+// Based on this doc: https://docs.nvidia.com/deploy/xid-errors/#topic_4
+var xidErrCodeToText = []string{
+	0:   "No Error",
+	1:   "Invalid or corrupted push buffer stream",
+	2:   "Invalid or corrupted push buffer stream",
+	3:   "Invalid or corrupted push buffer stream",
+	4:   "Invalid or corrupted push buffer stream",
+	5:   "Unused",
+	6:   "Invalid or corrupted push buffer stream",
+	7:   "Invalid or corrupted push buffer address",
+	8:   "GPU stopped processing",
+	9:   "Driver error programming GPU",
+	10:  "Unused",
+	11:  "Invalid or corrupted push buffer stream",
+	12:  "Driver error handling GPU exception",
+	13:  "Graphics Engine Exception",
+	14:  "Unused",
+	15:  "Unused",
+	16:  "Display engine hung",
+	17:  "Unused",
+	18:  "Bus mastering disabled in PCI Config Space",
+	19:  "Display Engine error",
+	20:  "Invalid or corrupted Mpeg push buffer",
+	21:  "Invalid or corrupted Motion Estimation push buffer",
+	22:  "Invalid or corrupted Video Processor push buffer",
+	23:  "Unused",
+	24:  "GPU semaphore timeout",
+	25:  "Invalid or illegal push buffer stream",
+	26:  "Framebuffer timeout",
+	27:  "Video processor exception",
+	28:  "Video processor exception",
+	29:  "Video processor exception",
+	30:  "GPU semaphore access error",
+	31:  "GPU memory page fault",
+	32:  "Invalid or corrupted push buffer stream",
+	33:  "Internal micro-controller error",
+	34:  "Video processor exception",
+	35:  "Video processor exception",
+	36:  "Video processor exception",
+	37:  "Driver firmware error",
+	38:  "Driver firmware error",
+	39:  "Unused",
+	40:  "Unused",
+	41:  "Unused",
+	42:  "Video processor exception",
+	43:  "GPU stopped processing",
+	44:  "Graphics Engine fault during context switch",
+	45:  "Preemptive cleanup, due to previous errors -- Most likely to see when running multiple cuda",
+	46:  "GPU stopped processing",
+	47:  "Video processor exception",
+	48:  "Double Bit ECC Error",
+	49:  "Unused",
+	50:  "Unused",
+	51:  "Unused",
+	52:  "Unused",
+	53:  "Unused",
+	54:  "Auxiliary power is not connected to the GPU board",
+	55:  "Unused",
+	56:  "Display Engine error",
+	57:  "Error programming video memory interface",
+	58:  "Unstable video memory interface detected",
+	59:  "Internal micro-controller error",
+	60:  "Video processor exception",
+	61:  "Internal micro-controller breakpoint/warning",
+	62:  "Internal micro-controller halt",
+	63:  "ECC page retirement or row remapping recording event",
+	64:  "ECC page retirement or row remapper recording failure",
+	65:  "Video processor exception",
+	66:  "Illegal access by driver",
+	67:  "Illegal access by driver",
+	68:  "NVDEC0 Exception",
+	69:  "Graphics Engine class error",
+	70:  "CE3: Unknown Error",
+	71:  "CE4: Unknown Error",
+	72:  "CE5: Unknown Error",
+	73:  "NVENC2 Error",
+	74:  "NVLINK Error",
+	75:  "CE6: Unknown Error",
+	76:  "CE7: Unknown Error",
+	77:  "CE8: Unknown Error",
+	78:  "vGPU Start Error",
+	79:  "GPU has fallen off the bus",
+	80:  "Corrupted data sent to GPU",
+	81:  "VGA Subsystem Error",
+	82:  "NVJPG0 Error",
+	83:  "NVDEC1 Error",
+	84:  "NVDEC2 Error",
+	85:  "CE9: Unknown Error",
+	86:  "OFA Exception",
+	87:  "Reserved",
+	88:  "NVDEC3 Error",
+	89:  "NVDEC4 Error",
+	90:  "Reserved",
+	91:  "Reserved",
+	92:  "High single-bit ECC error rate",
+	93:  "Non-fatal violation of provisioned InfoROM wear limit",
+	94:  "Contained ECC error",
+	95:  "Uncontained ECC error",
+	96:  "NVDEC5 Error",
+	97:  "NVDEC6 Error",
+	98:  "NVDEC7 Error",
+	99:  "NVJPG1 Error",
+	100: "NVJPG2 Error",
+	101: "NVJPG3 Error",
+	102: "NVJPG4 Error",
+	103: "NVJPG5 Error",
+	104: "NVJPG6 Error",
+	105: "NVJPG7 Error",
+	106: "SMBPBI Test Message",
+	107: "SMBPBI Test Message Silent",
+	108: "Reserved",
+	109: "Context Switch Timeout  Error",
+	110: "Security Fault Error",
+	111: "Display Bundle Error Event",
+	112: "Display Supervisor Error",
+	113: "DP Link Training Error",
+	114: "Display Pipeline Underflow Error",
+	115: "Display Core Channel Error",
+	116: "Display Window Channel Error",
+	117: "Display Cursor Channel Error",
+	118: "Display Pixel Pipeline Error",
+	119: "GSP RPC Timeout",
+	120: "GSP Error",
+	121: "C2C Link Error",
+	122: "SPI PMU RPC Read Failure",
+	123: "SPI PMU RPC Write Failure",
+	124: "SPI PMU RPC Erase Failure",
+	125: "Inforom FS Failure",
+	126: "Reserved",
+	127: "Reserved",
+	128: "Reserved",
+	129: "Reserved",
+	130: "Reserved",
+	131: "Reserved",
+	132: "Reserved",
+	133: "Reserved",
+	134: "Reserved",
+	135: "Reserved",
+	136: "Reserved",
+	137: "Reserved",
+	138: "Reserved",
+	139: "Reserved",
+	140: "Unrecovered ECC Error",
+	141: "Reserved",
+	142: "Reserved",
+	143: "GPU Initialization Failure",
+}