Merge latest development changes to main branch (#79)

* Cleanup: Remove unused code * Use Golang duration parser for 'interval' and 'duration' in main config * Update handling of LIKWID headers. Download only if not already present in the system. Fixes #73 * Units with cc-units (#64) * Add option to normalize units with cc-unit * Add unit conversion to router * Add option to change unit prefix in the router * Add to MetricRouter README * Add order of operations in router to README * Use second add_tags/del_tags only if metric gets renamed * Skip disks in DiskstatCollector that have size=0 * Check readability of sensor files in TempCollector * Fix for --once option * Rename `cpu` type to `hwthread` (#69) * Rename 'cpu' type to 'hwthread' to avoid naming clashes with MetricStore and CC-Webfrontend * Collectors in parallel (#74) * Provide info to CollectorManager whether the collector can be executed in parallel with others * Split serial and parallel collectors. Read in parallel first * Update NvidiaCollector with new metrics, MIG and NvLink support (#75) * CC topology module update (#76) * Rename CPU to hardware thread, write some comments * Do renaming in other parts * Remove CpuList and SocketList function from metricCollector. Available in ccTopology * Option to use MIG UUID as subtype-id in NvidiaCollector * Option to use MIG slice name as subtype-id in NvidiaCollector * MetricRouter: Fix JSON in README * Fix for Github Action to really use the selected version * Remove Ganglia installation in runonce Action and add Go 1.18 * Fix daemon options in init script * Add separate go.mod files to use it with deprecated 1.16 * Minor updates for Makefiles * fix string comparison * AMD ROCm SMI collector (#77) * Add collector for AMD ROCm SMI metrics * Fix import path * Fix imports * Remove Board Number * store GPU index explicitly * Remove board number from description * Use http instead of ftp to download likwid * Fix serial number in rocmCollector * Improved http sink (#78) * automatic flush in NatsSink * tweak default options of HttpSink * shorter cirt. section and retries for HttpSink * fix error handling * Remove file added by mistake. * Use http instead of ftp to download likwid * Fix serial number in rocmCollector Co-authored-by: Thomas Roehl <thomas.roehl@fau.de> Co-authored-by: Holger Obermaier <40787752+ho-ob@users.noreply.github.com> Co-authored-by: Lou <lou.knauer@gmx.de>
2025-07-29 16:16:06 +02:00 · 2022-06-08 15:25:40 +02:00
parent 186a62a86b
commit 8d85bd53f1
51 changed files with 2097 additions and 705 deletions
--- a/internal/ccTopology/ccTopology.go
+++ b/internal/ccTopology/ccTopology.go
@@ -29,6 +29,7 @@ func intArrayContains(array []int, str int) (int, bool) {
 	return -1, false
 }

+// Used internally for sysfs file reads
 func fileToInt(path string) int {
 	buffer, err := ioutil.ReadFile(path)
 	if err != nil {
@@ -47,6 +48,7 @@ func fileToInt(path string) int {
 	return int(id)
 }

+// Get list of CPU socket IDs
 func SocketList() []int {
 	buffer, err := ioutil.ReadFile(string(PROCFS_CPUINFO))
 	if err != nil {
@@ -54,7 +56,7 @@ func SocketList() []int {
 		return nil
 	}
 	ll := strings.Split(string(buffer), "\n")
-	var packs []int
+	packs := make([]int, 0)
 	for _, line := range ll {
 		if strings.HasPrefix(line, "physical id") {
 			lv := strings.Fields(line)
@@ -72,7 +74,8 @@ func SocketList() []int {
 	return packs
 }

-func CpuList() []int {
+// Get list of hardware thread IDs in the order of listing in /proc/cpuinfo
+func HwthreadList() []int {
 	buffer, err := ioutil.ReadFile(string(PROCFS_CPUINFO))
 	if err != nil {
 		log.Print(err)
@@ -97,6 +100,13 @@ func CpuList() []int {
 	return cpulist
 }

+// Get list of hardware thread IDs in the order of listing in /proc/cpuinfo
+// Deprecated! Use HwthreadList()
+func CpuList() []int {
+	return HwthreadList()
+}
+
+// Get list of CPU core IDs in the order of listing in /proc/cpuinfo
 func CoreList() []int {
 	buffer, err := ioutil.ReadFile(string(PROCFS_CPUINFO))
 	if err != nil {
@@ -122,6 +132,7 @@ func CoreList() []int {
 	return corelist
 }

+// Get list of NUMA node IDs
 func NumaNodeList() []int {
 	numaList := make([]int, 0)
 	globPath := filepath.Join(string(SYSFS_NUMABASE), "node*")
@@ -156,8 +167,9 @@ func NumaNodeList() []int {
 	return numaList
 }

+// Get list of CPU die IDs
 func DieList() []int {
-	cpulist := CpuList()
+	cpulist := HwthreadList()
 	dielist := make([]int, 0)
 	for _, c := range cpulist {
 		diepath := filepath.Join(string(SYSFS_CPUBASE), fmt.Sprintf("cpu%d", c), "topology/die_id")
@@ -175,7 +187,27 @@ func DieList() []int {
 	return SocketList()
 }

-type CpuEntry struct {
+// Get list of specified type using the naming format inside ClusterCockpit
+func GetTypeList(topology_type string) []int {
+	switch topology_type {
+	case "node":
+		return []int{0}
+	case "socket":
+		return SocketList()
+	case "die":
+		return DieList()
+	case "memoryDomain":
+		return NumaNodeList()
+	case "core":
+		return CoreList()
+	case "hwthread":
+		return HwthreadList()
+	}
+	return []int{}
+}
+
+// Structure holding all information about a hardware thread
+type HwthreadEntry struct {
 	Cpuid      int
 	SMT        int
 	Core       int
@@ -184,25 +216,25 @@ type CpuEntry struct {
 	Die        int
 }

-func CpuData() []CpuEntry {
+func CpuData() []HwthreadEntry {

-	fileToInt := func(path string) int {
-		buffer, err := ioutil.ReadFile(path)
-		if err != nil {
-			log.Print(err)
-			//cclogger.ComponentError("ccTopology", "Reading", path, ":", err.Error())
-			return -1
-		}
-		sbuffer := strings.Replace(string(buffer), "\n", "", -1)
-		var id int64
-		//_, err = fmt.Scanf("%d", sbuffer, &id)
-		id, err = strconv.ParseInt(sbuffer, 10, 32)
-		if err != nil {
-			cclogger.ComponentError("ccTopology", "Parsing", path, ":", sbuffer, err.Error())
-			return -1
-		}
-		return int(id)
-	}
+	// fileToInt := func(path string) int {
+	// 	buffer, err := ioutil.ReadFile(path)
+	// 	if err != nil {
+	// 		log.Print(err)
+	// 		//cclogger.ComponentError("ccTopology", "Reading", path, ":", err.Error())
+	// 		return -1
+	// 	}
+	// 	sbuffer := strings.Replace(string(buffer), "\n", "", -1)
+	// 	var id int64
+	// 	//_, err = fmt.Scanf("%d", sbuffer, &id)
+	// 	id, err = strconv.ParseInt(sbuffer, 10, 32)
+	// 	if err != nil {
+	// 		cclogger.ComponentError("ccTopology", "Parsing", path, ":", sbuffer, err.Error())
+	// 		return -1
+	// 	}
+	// 	return int(id)
+	// }
 	getCore := func(basepath string) int {
 		return fileToInt(fmt.Sprintf("%s/core_id", basepath))
 	}
@@ -260,9 +292,9 @@ func CpuData() []CpuEntry {
 		return 0
 	}

-	clist := make([]CpuEntry, 0)
-	for _, c := range CpuList() {
-		clist = append(clist, CpuEntry{Cpuid: c})
+	clist := make([]HwthreadEntry, 0)
+	for _, c := range HwthreadList() {
+		clist = append(clist, HwthreadEntry{Cpuid: c})
 	}
 	for i, centry := range clist {
 		centry.Socket = -1
@@ -298,6 +330,7 @@ func CpuData() []CpuEntry {
 	return clist
 }

+// Structure holding basic information about a CPU
 type CpuInformation struct {
 	NumHWthreads   int
 	SMTWidth       int
@@ -307,6 +340,7 @@ type CpuInformation struct {
 	NumNumaDomains int
 }

+// Get basic information about the CPU
 func CpuInfo() CpuInformation {
 	var c CpuInformation

@@ -342,7 +376,8 @@ func CpuInfo() CpuInformation {
 	return c
 }

-func GetCpuSocket(cpuid int) int {
+// Get the CPU socket ID for a given hardware thread ID
+func GetHwthreadSocket(cpuid int) int {
 	cdata := CpuData()
 	for _, d := range cdata {
 		if d.Cpuid == cpuid {
@@ -352,7 +387,8 @@ func GetCpuSocket(cpuid int) int {
 	return -1
 }

-func GetCpuNumaDomain(cpuid int) int {
+// Get the NUMA node ID for a given hardware thread ID
+func GetHwthreadNumaDomain(cpuid int) int {
 	cdata := CpuData()
 	for _, d := range cdata {
 		if d.Cpuid == cpuid {
@@ -362,7 +398,8 @@ func GetCpuNumaDomain(cpuid int) int {
 	return -1
 }

-func GetCpuDie(cpuid int) int {
+// Get the CPU die ID for a given hardware thread ID
+func GetHwthreadDie(cpuid int) int {
 	cdata := CpuData()
 	for _, d := range cdata {
 		if d.Cpuid == cpuid {
@@ -372,7 +409,8 @@ func GetCpuDie(cpuid int) int {
 	return -1
 }

-func GetCpuCore(cpuid int) int {
+// Get the CPU core ID for a given hardware thread ID
+func GetHwthreadCore(cpuid int) int {
 	cdata := CpuData()
 	for _, d := range cdata {
 		if d.Cpuid == cpuid {
@@ -382,7 +420,8 @@ func GetCpuCore(cpuid int) int {
 	return -1
 }

-func GetSocketCpus(socket int) []int {
+// Get the all hardware thread ID associated with a CPU socket
+func GetSocketHwthreads(socket int) []int {
 	all := CpuData()
 	cpulist := make([]int, 0)
 	for _, d := range all {
@@ -393,7 +432,8 @@ func GetSocketCpus(socket int) []int {
 	return cpulist
 }

-func GetNumaDomainCpus(domain int) []int {
+// Get the all hardware thread ID associated with a NUMA node
+func GetNumaDomainHwthreads(domain int) []int {
 	all := CpuData()
 	cpulist := make([]int, 0)
 	for _, d := range all {
@@ -404,7 +444,8 @@ func GetNumaDomainCpus(domain int) []int {
 	return cpulist
 }

-func GetDieCpus(die int) []int {
+// Get the all hardware thread ID associated with a CPU die
+func GetDieHwthreads(die int) []int {
 	all := CpuData()
 	cpulist := make([]int, 0)
 	for _, d := range all {
@@ -415,7 +456,8 @@ func GetDieCpus(die int) []int {
 	return cpulist
 }

-func GetCoreCpus(core int) []int {
+// Get the all hardware thread ID associated with a CPU core
+func GetCoreHwthreads(core int) []int {
 	all := CpuData()
 	cpulist := make([]int, 0)
 	for _, d := range all {
--- a/internal/metricAggregator/metricAggregatorFunctions.go
+++ b/internal/metricAggregator/metricAggregatorFunctions.go
@@ -246,7 +246,7 @@ func matchfunc(args ...interface{}) (interface{}, error) {
 func getCpuCoreFunc(args ...interface{}) (interface{}, error) {
 	switch cpuid := args[0].(type) {
 	case int:
-		return topo.GetCpuCore(cpuid), nil
+		return topo.GetHwthreadCore(cpuid), nil
 	}
 	return -1, errors.New("function 'getCpuCore' accepts only an 'int' cpuid")
 }
@@ -255,7 +255,7 @@ func getCpuCoreFunc(args ...interface{}) (interface{}, error) {
 func getCpuSocketFunc(args ...interface{}) (interface{}, error) {
 	switch cpuid := args[0].(type) {
 	case int:
-		return topo.GetCpuSocket(cpuid), nil
+		return topo.GetHwthreadSocket(cpuid), nil
 	}
 	return -1, errors.New("function 'getCpuCore' accepts only an 'int' cpuid")
 }
@@ -264,7 +264,7 @@ func getCpuSocketFunc(args ...interface{}) (interface{}, error) {
 func getCpuNumaDomainFunc(args ...interface{}) (interface{}, error) {
 	switch cpuid := args[0].(type) {
 	case int:
-		return topo.GetCpuNumaDomain(cpuid), nil
+		return topo.GetHwthreadNumaDomain(cpuid), nil
 	}
 	return -1, errors.New("function 'getCpuNuma' accepts only an 'int' cpuid")
 }
@@ -273,7 +273,7 @@ func getCpuNumaDomainFunc(args ...interface{}) (interface{}, error) {
 func getCpuDieFunc(args ...interface{}) (interface{}, error) {
 	switch cpuid := args[0].(type) {
 	case int:
-		return topo.GetCpuDie(cpuid), nil
+		return topo.GetHwthreadDie(cpuid), nil
 	}
 	return -1, errors.New("function 'getCpuDie' accepts only an 'int' cpuid")
 }
@@ -336,7 +336,7 @@ func getCpuListOfDieFunc(args ...interface{}) (interface{}, error) {

 // wrapper function to get a list of all cpuids of the node
 func getCpuListOfNode(args ...interface{}) (interface{}, error) {
-	return topo.CpuList(), nil
+	return topo.HwthreadList(), nil
 }

 // helper function to get the cpuid list for a CCMetric type tag set (type and type-id)
@@ -348,14 +348,14 @@ func getCpuListOfType(args ...interface{}) (interface{}, error) {
 	case string:
 		switch typ {
 		case "node":
-			return topo.CpuList(), nil
+			return topo.HwthreadList(), nil
 		case "socket":
 			return getCpuListOfSocketFunc(args[1])
 		case "numadomain":
 			return getCpuListOfNumaDomainFunc(args[1])
 		case "core":
 			return getCpuListOfCoreFunc(args[1])
-		case "cpu":
+		case "hwthread":
 			var cpu int

 			switch id := args[1].(type) {
--- a/internal/metricRouter/README.md
+++ b/internal/metricRouter/README.md
@@ -52,6 +52,11 @@ The CCMetric router sits in between the collectors and the sinks and can be used
    ],
    "rename_metrics" : {
        "metric_12345" : "mymetric"
+    },
+    "normalize_units" : true,
+    "change_unit_prefix" : {
+      "mem_used" : "G",
+      "mem_total" : "G"
    }
 }
 ```
@@ -192,6 +197,14 @@ This option takes a list of evaluable conditions and performs them one after the
 ```
 The first line is comparable with the example in `drop_metrics`, it drops all metrics starting with `drop_metric_` and ending with a number. The second line drops all metrics of the first hardware thread (**not** recommended)

+# Manipulating the metric units
+
+## The `normalize_units` option
+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-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
+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

@@ -239,3 +252,22 @@ Use cases for `interval_aggregates`:
    }
  }
 ```
+
+# Order of operations
+
+The router performs the above mentioned options in a specific order. In order to get the logic you want for a specific metric, it is crucial to know the processing order:
+
+- Add the `hostname` tag (c)
+- Manipulate the timestamp to the interval timestamp (c,r)
+- Drop metrics based on `drop_metrics` and `drop_metrics_if` (c,r)
+- Add tags based on `add_tags` (c,r)
+- Delete tags based on `del_tags` (c,r)
+- Rename metric based on `rename_metric` (c,r)
+  - Add tags based on `add_tags` to still work if the configuration uses the new name (c,r) 
+  - Delete tags based on `del_tags` to still work if the configuration uses the new name (c,r)
+- Normalize units when `normalize_units` is set (c,r)
+- Convert unit prefix based on `change_unit_prefix` (c,r)
+
+Legend:
+- 'c' if metric is coming from a collector
+- 'r' if metric is coming from a receiver
--- a/internal/metricRouter/metricRouter.go
+++ b/internal/metricRouter/metricRouter.go
@@ -12,6 +12,7 @@ import (
 	lp "github.com/ClusterCockpit/cc-metric-collector/internal/ccMetric"
 	agg "github.com/ClusterCockpit/cc-metric-collector/internal/metricAggregator"
 	mct "github.com/ClusterCockpit/cc-metric-collector/internal/multiChanTicker"
+	units "github.com/ClusterCockpit/cc-units"
 )

 const ROUTER_MAX_FORWARD = 50
@@ -35,6 +36,8 @@ type metricRouterConfig struct {
 	IntervalStamp     bool                                 `json:"interval_timestamp"`  // Update timestamp periodically by ticker each interval?
 	NumCacheIntervals int                                  `json:"num_cache_intervals"` // Number of intervals of cached metrics for evaluation
 	MaxForward        int                                  `json:"max_forward"`         // Number of maximal forwarded metrics at one select
+	NormalizeUnits    bool                                 `json:"normalize_units"`     // Check unit meta flag and normalize it using cc-units
+	ChangeUnitPrefix  map[string]string                    `json:"change_unit_prefix"`  // Add prefix that should be applied to the metrics
 	dropMetrics       map[string]bool                      // Internal map for O(1) lookup
 }

@@ -207,6 +210,38 @@ func (r *metricRouter) dropMetric(point lp.CCMetric) bool {
 	return false
 }

+func (r *metricRouter) prepareUnit(point lp.CCMetric) bool {
+	if r.config.NormalizeUnits {
+		if in_unit, ok := point.GetMeta("unit"); ok {
+			u := units.NewUnit(in_unit)
+			if u.Valid() {
+				point.AddMeta("unit", u.Short())
+			}
+		}
+	}
+	if newP, ok := r.config.ChangeUnitPrefix[point.Name()]; ok {
+
+		newPrefix := units.NewPrefix(newP)
+
+		if in_unit, ok := point.GetMeta("unit"); ok && newPrefix != units.InvalidPrefix {
+			u := units.NewUnit(in_unit)
+			if u.Valid() {
+				cclog.ComponentDebug("MetricRouter", "Change prefix to", newP, "for metric", point.Name())
+				conv, out_unit := units.GetUnitPrefixFactor(u, newPrefix)
+				if conv != nil && out_unit.Valid() {
+					if val, ok := point.GetField("value"); ok {
+						point.AddField("value", conv(val))
+						point.AddMeta("unit", out_unit.Short())
+					}
+				}
+			}
+
+		}
+	}
+
+	return true
+}
+
 // Start starts the metric router
 func (r *metricRouter) Start() {
 	// start timer if configured
@@ -232,9 +267,11 @@ func (r *metricRouter) Start() {
 		if new, ok := r.config.RenameMetrics[name]; ok {
 			point.SetName(new)
 			point.AddMeta("oldname", name)
+			r.DoAddTags(point)
+			r.DoDelTags(point)
 		}
-		r.DoAddTags(point)
-		r.DoDelTags(point)
+
+		r.prepareUnit(point)

 		for _, o := range r.outputs {
 			o <- point