Update Hugo integration

collectors/nvidiaMetric.go

@@ -31,8 +31,6 @@ type NvidiaCollectorDevice struct {
 	excludeMetrics map[string]bool
 	tags           map[string]string
 	meta           map[string]string
-	lastEnergyReading   uint64
-	lastEnergyTimestamp time.Time
 }
 
 type NvidiaCollector struct {
@@ -151,8 +149,6 @@ func (m *NvidiaCollector) Init(config json.RawMessage) error {
 
 		// Add device handle
 		g.device = device
-		g.lastEnergyReading = 0
-		g.lastEnergyTimestamp = time.Now()
 
 		// Add tags
 		g.tags = map[string]string{
@@ -210,7 +206,7 @@ func (m *NvidiaCollector) Init(config json.RawMessage) error {
 	return nil
 }
 
-func readMemoryInfo(device *NvidiaCollectorDevice, output chan lp.CCMessage) error {
+func readMemoryInfo(device NvidiaCollectorDevice, output chan lp.CCMessage) error {
 	if !device.excludeMetrics["nv_fb_mem_total"] || !device.excludeMetrics["nv_fb_mem_used"] || !device.excludeMetrics["nv_fb_mem_reserved"] {
 		var total uint64
 		var used uint64
@@ -254,7 +250,7 @@ func readMemoryInfo(device *NvidiaCollectorDevice, output chan lp.CCMessage) err
 	return nil
 }
 
-func readBarMemoryInfo(device *NvidiaCollectorDevice, output chan lp.CCMessage) error {
+func readBarMemoryInfo(device NvidiaCollectorDevice, output chan lp.CCMessage) error {
 	if !device.excludeMetrics["nv_bar1_mem_total"] || !device.excludeMetrics["nv_bar1_mem_used"] {
 		meminfo, ret := nvml.DeviceGetBAR1MemoryInfo(device.device)
 		if ret != nvml.SUCCESS {
@@ -281,7 +277,7 @@ func readBarMemoryInfo(device *NvidiaCollectorDevice, output chan lp.CCMessage)
 	return nil
 }
 
-func readUtilization(device *NvidiaCollectorDevice, output chan lp.CCMessage) error {
+func readUtilization(device NvidiaCollectorDevice, output chan lp.CCMessage) error {
 	isMig, ret := nvml.DeviceIsMigDeviceHandle(device.device)
 	if ret != nvml.SUCCESS {
 		err := errors.New(nvml.ErrorString(ret))
@@ -323,7 +319,7 @@ func readUtilization(device *NvidiaCollectorDevice, output chan lp.CCMessage) er
 	return nil
 }
 
-func readTemp(device *NvidiaCollectorDevice, output chan lp.CCMessage) error {
+func readTemp(device NvidiaCollectorDevice, output chan lp.CCMessage) error {
 	if !device.excludeMetrics["nv_temp"] {
 		// Retrieves the current temperature readings for the device, in degrees C.
 		//
@@ -342,7 +338,7 @@ func readTemp(device *NvidiaCollectorDevice, output chan lp.CCMessage) error {
 	return nil
 }
 
-func readFan(device *NvidiaCollectorDevice, output chan lp.CCMessage) error {
+func readFan(device NvidiaCollectorDevice, output chan lp.CCMessage) error {
 	if !device.excludeMetrics["nv_fan"] {
 		// Retrieves the intended operating speed of the device's fan.
 		//
@@ -365,7 +361,7 @@ func readFan(device *NvidiaCollectorDevice, output chan lp.CCMessage) error {
 	return nil
 }
 
-// func readFans(device *NvidiaCollectorDevice, output chan lp.CCMessage) error {
+// func readFans(device NvidiaCollectorDevice, output chan lp.CCMessage) error {
 // 	if !device.excludeMetrics["nv_fan"] {
 // 		numFans, ret := nvml.DeviceGetNumFans(device.device)
 // 		if ret == nvml.SUCCESS {
@@ -386,7 +382,7 @@ func readFan(device *NvidiaCollectorDevice, output chan lp.CCMessage) error {
 // 	return nil
 // }
 
-func readEccMode(device *NvidiaCollectorDevice, output chan lp.CCMessage) error {
+func readEccMode(device NvidiaCollectorDevice, output chan lp.CCMessage) error {
 	if !device.excludeMetrics["nv_ecc_mode"] {
 		// Retrieves the current and pending ECC modes for the device.
 		//
@@ -420,7 +416,7 @@ func readEccMode(device *NvidiaCollectorDevice, output chan lp.CCMessage) error
 	return nil
 }
 
-func readPerfState(device *NvidiaCollectorDevice, output chan lp.CCMessage) error {
+func readPerfState(device NvidiaCollectorDevice, output chan lp.CCMessage) error {
 	if !device.excludeMetrics["nv_perf_state"] {
 		// Retrieves the current performance state for the device.
 		//
@@ -440,16 +436,13 @@ func readPerfState(device *NvidiaCollectorDevice, output chan lp.CCMessage) erro
 	return nil
 }
 
-func readPowerUsage(device *NvidiaCollectorDevice, output chan lp.CCMessage) error {
+func readPowerUsage(device NvidiaCollectorDevice, output chan lp.CCMessage) error {
 	if !device.excludeMetrics["nv_power_usage"] {
 		// Retrieves power usage for this GPU in milliwatts and its associated circuitry (e.g. memory)
 		//
 		// On Fermi and Kepler GPUs the reading is accurate to within +/- 5% of current power draw.
-		// On Ampere (except GA100) or newer GPUs, the API returns power averaged over 1 sec interval.
-		// On GA100 and older architectures, instantaneous power is returned.
 		//
-		// It is only available if power management mode is supported.
-
+		// It is only available if power management mode is supported
 		mode, ret := nvml.DeviceGetPowerManagementMode(device.device)
 		if ret != nvml.SUCCESS {
 			return nil
@@ -468,54 +461,7 @@ func readPowerUsage(device *NvidiaCollectorDevice, output chan lp.CCMessage) err
 	return nil
 }
 
-func readEnergyConsumption(device *NvidiaCollectorDevice, output chan lp.CCMessage) error {
-	// Retrieves total energy consumption for this GPU in millijoules (mJ) since the driver was last reloaded
-
-	// For Volta or newer fully supported devices.
-	if (!device.excludeMetrics["nv_energy"]) && (!device.excludeMetrics["nv_energy_abs"]) && (!device.excludeMetrics["nv_average_power"]) {
-		now := time.Now()
-		mode, ret := nvml.DeviceGetPowerManagementMode(device.device)
-		if ret != nvml.SUCCESS {
-			return nil
-		}
-		if mode == nvml.FEATURE_ENABLED {
-			energy, ret := nvml.DeviceGetTotalEnergyConsumption(device.device)
-			if ret == nvml.SUCCESS {
-				if device.lastEnergyReading != 0 {
-					if !device.excludeMetrics["nv_energy"] {
-						y, err := lp.NewMetric("nv_energy", device.tags, device.meta, (energy-device.lastEnergyReading)/1000, now)
-						if err == nil {
-							y.AddMeta("unit", "Joules")
-							output <- y
-						}
-					}
-					if !device.excludeMetrics["nv_average_power"] {
-
-						energyDiff := (energy - device.lastEnergyReading) / 1000
-						timeDiff := now.Sub(device.lastEnergyTimestamp)
-						y, err := lp.NewMetric("nv_average_power", device.tags, device.meta, energyDiff/uint64(timeDiff.Seconds()), now)
-						if err == nil {
-							y.AddMeta("unit", "watts")
-							output <- y
-						}
-					}
-				}
-				if !device.excludeMetrics["nv_energy_abs"] {
-					y, err := lp.NewMetric("nv_energy_abs", device.tags, device.meta, energy/1000, now)
-					if err == nil {
-						y.AddMeta("unit", "Joules")
-						output <- y
-					}
-				}
-				device.lastEnergyReading = energy
-				device.lastEnergyTimestamp = time.Now()
-			}
-		}
-	}
-	return nil
-}
-
-func readClocks(device *NvidiaCollectorDevice, output chan lp.CCMessage) error {
+func readClocks(device NvidiaCollectorDevice, output chan lp.CCMessage) error {
 	// Retrieves the current clock speeds for the device.
 	//
 	// Available clock information:
@@ -567,7 +513,7 @@ func readClocks(device *NvidiaCollectorDevice, output chan lp.CCMessage) error {
 	return nil
 }
 
-func readMaxClocks(device *NvidiaCollectorDevice, output chan lp.CCMessage) error {
+func readMaxClocks(device NvidiaCollectorDevice, output chan lp.CCMessage) error {
 	// Retrieves the maximum clock speeds for the device.
 	//
 	// Available clock information:
@@ -625,7 +571,7 @@ func readMaxClocks(device *NvidiaCollectorDevice, output chan lp.CCMessage) erro
 	return nil
 }
 
-func readEccErrors(device *NvidiaCollectorDevice, output chan lp.CCMessage) error {
+func readEccErrors(device NvidiaCollectorDevice, output chan lp.CCMessage) error {
 	if !device.excludeMetrics["nv_ecc_uncorrected_error"] {
 		// Retrieves the total ECC error counts for the device.
 		//
@@ -656,7 +602,7 @@ func readEccErrors(device *NvidiaCollectorDevice, output chan lp.CCMessage) erro
 	return nil
 }
 
-func readPowerLimit(device *NvidiaCollectorDevice, output chan lp.CCMessage) error {
+func readPowerLimit(device NvidiaCollectorDevice, output chan lp.CCMessage) error {
 	if !device.excludeMetrics["nv_power_max_limit"] {
 		// Retrieves the power management limit associated with this device.
 		//
@@ -676,7 +622,7 @@ func readPowerLimit(device *NvidiaCollectorDevice, output chan lp.CCMessage) err
 	return nil
 }
 
-func readEncUtilization(device *NvidiaCollectorDevice, output chan lp.CCMessage) error {
+func readEncUtilization(device NvidiaCollectorDevice, output chan lp.CCMessage) error {
 	isMig, ret := nvml.DeviceIsMigDeviceHandle(device.device)
 	if ret != nvml.SUCCESS {
 		err := errors.New(nvml.ErrorString(ret))
@@ -703,7 +649,7 @@ func readEncUtilization(device *NvidiaCollectorDevice, output chan lp.CCMessage)
 	return nil
 }
 
-func readDecUtilization(device *NvidiaCollectorDevice, output chan lp.CCMessage) error {
+func readDecUtilization(device NvidiaCollectorDevice, output chan lp.CCMessage) error {
 	isMig, ret := nvml.DeviceIsMigDeviceHandle(device.device)
 	if ret != nvml.SUCCESS {
 		err := errors.New(nvml.ErrorString(ret))
@@ -730,7 +676,7 @@ func readDecUtilization(device *NvidiaCollectorDevice, output chan lp.CCMessage)
 	return nil
 }
 
-func readRemappedRows(device *NvidiaCollectorDevice, output chan lp.CCMessage) error {
+func readRemappedRows(device NvidiaCollectorDevice, output chan lp.CCMessage) error {
 	if !device.excludeMetrics["nv_remapped_rows_corrected"] ||
 		!device.excludeMetrics["nv_remapped_rows_uncorrected"] ||
 		!device.excludeMetrics["nv_remapped_rows_pending"] ||
@@ -783,7 +729,7 @@ func readRemappedRows(device *NvidiaCollectorDevice, output chan lp.CCMessage) e
 	return nil
 }
 
-func readProcessCounts(device *NvidiaCollectorDevice, output chan lp.CCMessage) error {
+func readProcessCounts(device NvidiaCollectorDevice, output chan lp.CCMessage) error {
 	if !device.excludeMetrics["nv_compute_processes"] {
 		// Get information about processes with a compute context on a device
 		//
@@ -875,7 +821,7 @@ func readProcessCounts(device *NvidiaCollectorDevice, output chan lp.CCMessage)
 	return nil
 }
 
-func readViolationStats(device *NvidiaCollectorDevice, output chan lp.CCMessage) error {
+func readViolationStats(device NvidiaCollectorDevice, output chan lp.CCMessage) error {
 	var violTime nvml.ViolationTime
 	var ret nvml.Return
 
@@ -989,7 +935,7 @@ func readViolationStats(device *NvidiaCollectorDevice, output chan lp.CCMessage)
 	return nil
 }
 
-func readNVLinkStats(device *NvidiaCollectorDevice, output chan lp.CCMessage) error {
+func readNVLinkStats(device NvidiaCollectorDevice, output chan lp.CCMessage) error {
 	// Retrieves the specified error counter value
 	// Please refer to \a nvmlNvLinkErrorCounter_t for error counters that are available
 	//
@@ -1124,7 +1070,7 @@ func (m *NvidiaCollector) Read(interval time.Duration, output chan lp.CCMessage)
 		return
 	}
 
-	readAll := func(device *NvidiaCollectorDevice, output chan lp.CCMessage) {
+	readAll := func(device NvidiaCollectorDevice, output chan lp.CCMessage) {
 		name, ret := nvml.DeviceGetName(device.device)
 		if ret != nvml.SUCCESS {
 			name = "NoName"
@@ -1164,11 +1110,6 @@ func (m *NvidiaCollector) Read(interval time.Duration, output chan lp.CCMessage)
 			cclog.ComponentDebug(m.name, "readPowerUsage for device", name, "failed")
 		}
 
-		err = readEnergyConsumption(device, output)
-		if err != nil {
-			cclog.ComponentDebug(m.name, "readEnergyConsumption for device", name, "failed")
-		}
-
 		err = readClocks(device, output)
 		if err != nil {
 			cclog.ComponentDebug(m.name, "readClocks for device", name, "failed")
@@ -1228,7 +1169,7 @@ func (m *NvidiaCollector) Read(interval time.Duration, output chan lp.CCMessage)
 	// Actual read loop over all attached Nvidia GPUs
 	for i := 0; i < m.num_gpus; i++ {
 
-		readAll(&m.gpus[i], output)
+		readAll(m.gpus[i], output)
 
 		// Iterate over all MIG devices if any
 		if m.config.ProcessMigDevices {
@@ -1302,7 +1243,7 @@ func (m *NvidiaCollector) Read(interval time.Duration, output chan lp.CCMessage)
 					}
 				}
 
-				readAll(&migDevice, output)
+				readAll(migDevice, output)
 			}
 		}
 	}

collectors/nvidiaMetric.md

@@ -82,8 +82,5 @@ Metrics:
 * `nv_nvlink_ecc_errors`
 * `nv_nvlink_replay_errors`
 * `nv_nvlink_recovery_errors`
-* `nv_energy`
-* `nv_energy_abs`
-* `nv_average_power`
 
 Some metrics add the additional sub type tag (`stype`) like the `nv_nvlink_*` metrics set `stype=nvlink,stype-id=<link_number>`. 

internal/metricRouter/README.md

@@ -236,13 +236,13 @@ __deprecated__
 
 
 The cc-metric-collector tries to read the data from the system as it is reported. If available, it tries to read the metric unit from the system as well (e.g. from `/proc/meminfo`). The problem is that, depending on the source, the metric units are named differently. Just think about `byte`, `Byte`, `B`, `bytes`, ...
-The [cc-units](https://github.com/ClusterCockpit/cc-lib/ccUnits) package provides us a normalization option to use the same metric unit name for all metrics. It this option is set to true, all `unit` meta tags are normalized.
+The [cc-units](https://github.com/ClusterCockpit/cc-units) package provides us a normalization option to use the same metric unit name for all metrics. It this option is set to true, all `unit` meta tags are normalized.
 
 ## The `change_unit_prefix` section
 
 __deprecated__
 
-It is often the case that metrics are reported by the system using a rather outdated unit prefix (like `/proc/meminfo` still uses kByte despite current memory sizes are in the GByte range). If you want to change the prefix of a unit, you can do that with the help of [cc-units](https://github.com/ClusterCockpit/cc-lib/ccUnits). The setting works on the metric name and requires the new prefix for the metric. The cc-units package determines the scaling factor.
+It is often the case that metrics are reported by the system using a rather outdated unit prefix (like `/proc/meminfo` still uses kByte despite current memory sizes are in the GByte range). If you want to change the prefix of a unit, you can do that with the help of [cc-units](https://github.com/ClusterCockpit/cc-units). The setting works on the metric name and requires the new prefix for the metric. The cc-units package determines the scaling factor.
 
 # Aggregate metric values of the current interval with the `interval_aggregates` option