diff --git a/collectors/cpufreqCpuinfoMetric.go b/collectors/cpufreqCpuinfoMetric.go
index 44a3b0c..6c3de7a 100644
--- a/collectors/cpufreqCpuinfoMetric.go
+++ b/collectors/cpufreqCpuinfoMetric.go
@@ -57,7 +57,7 @@ func (m *CPUFreqCpuInfoCollector) Init(config json.RawMessage) error {
const cpuInfoFile = "/proc/cpuinfo"
file, err := os.Open(cpuInfoFile)
if err != nil {
- return fmt.Errorf("Failed to open file '%s': %v", cpuInfoFile, err)
+ return fmt.Errorf("failed to open file '%s': %v", cpuInfoFile, err)
}
defer file.Close()
@@ -106,14 +106,14 @@ func (m *CPUFreqCpuInfoCollector) Init(config json.RawMessage) error {
topology.coreID = coreID
topology.coreID_int, err = strconv.ParseInt(coreID, 10, 64)
if err != nil {
- return fmt.Errorf("Unable to convert coreID '%s' to int64: %v", coreID, err)
+ return fmt.Errorf("unable to convert coreID '%s' to int64: %v", coreID, err)
}
// Physical package ID
topology.physicalPackageID = physicalPackageID
topology.physicalPackageID_int, err = strconv.ParseInt(physicalPackageID, 10, 64)
if err != nil {
- return fmt.Errorf("Unable to convert physicalPackageID '%s' to int64: %v", physicalPackageID, err)
+ return fmt.Errorf("unable to convert physicalPackageID '%s' to int64: %v", physicalPackageID, err)
}
// increase maximun socket / package ID, when required
diff --git a/collectors/cpufreqMetric.go b/collectors/cpufreqMetric.go
index 5146baa..0bf6d4c 100644
--- a/collectors/cpufreqMetric.go
+++ b/collectors/cpufreqMetric.go
@@ -70,10 +70,10 @@ func (m *CPUFreqCollector) Init(config json.RawMessage) error {
globPattern := filepath.Join(baseDir, "cpu[0-9]*")
cpuDirs, err := filepath.Glob(globPattern)
if err != nil {
- return fmt.Errorf("Unable to glob files with pattern '%s': %v", globPattern, err)
+ return fmt.Errorf("unable to glob files with pattern '%s': %v", globPattern, err)
}
if cpuDirs == nil {
- return fmt.Errorf("Unable to find any files with pattern '%s'", globPattern)
+ return fmt.Errorf("unable to find any files with pattern '%s'", globPattern)
}
// Initialize CPU topology
@@ -82,38 +82,38 @@ func (m *CPUFreqCollector) Init(config json.RawMessage) error {
processor := strings.TrimPrefix(cpuDir, "/sys/devices/system/cpu/cpu")
processor_int, err := strconv.ParseInt(processor, 10, 64)
if err != nil {
- return fmt.Errorf("Unable to convert cpuID '%s' to int64: %v", processor, err)
+ return fmt.Errorf("unable to convert cpuID '%s' to int64: %v", processor, err)
}
// Read package ID
physicalPackageIDFile := filepath.Join(cpuDir, "topology", "physical_package_id")
line, err := ioutil.ReadFile(physicalPackageIDFile)
if err != nil {
- return fmt.Errorf("Unable to read physical package ID from file '%s': %v", physicalPackageIDFile, err)
+ return fmt.Errorf("unable to read physical package ID from file '%s': %v", physicalPackageIDFile, err)
}
physicalPackageID := strings.TrimSpace(string(line))
physicalPackageID_int, err := strconv.ParseInt(physicalPackageID, 10, 64)
if err != nil {
- return fmt.Errorf("Unable to convert packageID '%s' to int64: %v", physicalPackageID, err)
+ return fmt.Errorf("unable to convert packageID '%s' to int64: %v", physicalPackageID, err)
}
// Read core ID
coreIDFile := filepath.Join(cpuDir, "topology", "core_id")
line, err = ioutil.ReadFile(coreIDFile)
if err != nil {
- return fmt.Errorf("Unable to read core ID from file '%s': %v", coreIDFile, err)
+ return fmt.Errorf("unable to read core ID from file '%s': %v", coreIDFile, err)
}
coreID := strings.TrimSpace(string(line))
coreID_int, err := strconv.ParseInt(coreID, 10, 64)
if err != nil {
- return fmt.Errorf("Unable to convert coreID '%s' to int64: %v", coreID, err)
+ return fmt.Errorf("unable to convert coreID '%s' to int64: %v", coreID, err)
}
// Check access to current frequency file
scalingCurFreqFile := filepath.Join(cpuDir, "cpufreq", "scaling_cur_freq")
err = unix.Access(scalingCurFreqFile, unix.R_OK)
if err != nil {
- return fmt.Errorf("Unable to access file '%s': %v", scalingCurFreqFile, err)
+ return fmt.Errorf("unable to access file '%s': %v", scalingCurFreqFile, err)
}
t := &m.topology[processor_int]
diff --git a/collectors/cpustatMetric.go b/collectors/cpustatMetric.go
index 28ae002..556aad4 100644
--- a/collectors/cpustatMetric.go
+++ b/collectors/cpustatMetric.go
@@ -21,11 +21,10 @@ type CpustatCollectorConfig struct {
type CpustatCollector struct {
metricCollector
- config CpustatCollectorConfig
- matches map[string]int
- cputags map[string]map[string]string
- nodetags map[string]string
- num_cpus_metric lp.CCMetric
+ config CpustatCollectorConfig
+ matches map[string]int
+ cputags map[string]map[string]string
+ nodetags map[string]string
}
func (m *CpustatCollector) Init(config json.RawMessage) error {
diff --git a/collectors/customCmdMetric.go b/collectors/customCmdMetric.go
index e978c49..ec2109b 100644
--- a/collectors/customCmdMetric.go
+++ b/collectors/customCmdMetric.go
@@ -61,7 +61,7 @@ func (m *CustomCmdCollector) Init(config json.RawMessage) error {
}
}
if len(m.files) == 0 && len(m.commands) == 0 {
- return errors.New("No metrics to collect")
+ return errors.New("no metrics to collect")
}
m.handler = influx.NewMetricHandler()
m.parser = influx.NewParser(m.handler)
diff --git a/collectors/diskstatMetric.go b/collectors/diskstatMetric.go
index 16c70ba..4910c83 100644
--- a/collectors/diskstatMetric.go
+++ b/collectors/diskstatMetric.go
@@ -3,7 +3,6 @@ package collectors
import (
"bufio"
"encoding/json"
- "fmt"
"os"
"strings"
"syscall"
@@ -81,8 +80,7 @@ func (m *DiskstatCollector) Read(interval time.Duration, output chan lp.CCMetric
stat := syscall.Statfs_t{}
err := syscall.Statfs(path, &stat)
if err != nil {
- fmt.Println(err.Error())
- return
+ continue
}
tags := map[string]string{"type": "node", "device": linefields[0]}
total := (stat.Blocks * uint64(stat.Bsize)) / uint64(1000000000)
diff --git a/collectors/gpfsMetric.go b/collectors/gpfsMetric.go
index 26fc723..ed63201 100644
--- a/collectors/gpfsMetric.go
+++ b/collectors/gpfsMetric.go
@@ -70,6 +70,7 @@ func (m *GpfsCollector) Init(config json.RawMessage) error {
for _, fs := range m.config.ExcludeFilesystem {
m.skipFS[fs] = struct{}{}
}
+ m.lastState = make(map[string]GpfsCollectorLastState)
// GPFS / IBM Spectrum Scale file system statistics can only be queried by user root
user, err := user.Current()
@@ -162,11 +163,16 @@ func (m *GpfsCollector) Read(interval time.Duration, output chan lp.CCMetric) {
continue
}
+ // Add filesystem tag
m.tags["filesystem"] = filesystem
- if _, ok := m.lastState[filesystem]; !ok {
- m.lastState[filesystem] = GpfsCollectorLastState{
- bytesRead: -1,
- bytesWritten: -1,
+
+ // Create initial last state
+ if m.config.SendBandwidths {
+ if _, ok := m.lastState[filesystem]; !ok {
+ m.lastState[filesystem] = GpfsCollectorLastState{
+ bytesRead: -1,
+ bytesWritten: -1,
+ }
}
}
diff --git a/collectors/infinibandMetric.go b/collectors/infinibandMetric.go
index 5be095d..274e669 100644
--- a/collectors/infinibandMetric.go
+++ b/collectors/infinibandMetric.go
@@ -18,13 +18,18 @@ import (
const IB_BASEPATH = "/sys/class/infiniband/"
+type InfinibandCollectorMetric struct {
+ path string
+ unit string
+}
+
type InfinibandCollectorInfo struct {
- LID string // IB local Identifier (LID)
- device string // IB device
- port string // IB device port
- portCounterFiles map[string]string // mapping counter name -> sysfs file
- tagSet map[string]string // corresponding tag list
- lastState map[string]int64 // State from last measurement
+ LID string // IB local Identifier (LID)
+ device string // IB device
+ port string // IB device port
+ portCounterFiles map[string]InfinibandCollectorMetric // mapping counter name -> InfinibandCollectorMetric
+ tagSet map[string]string // corresponding tag list
+ lastState map[string]int64 // State from last measurement
}
type InfinibandCollector struct {
@@ -106,16 +111,16 @@ func (m *InfinibandCollector) Init(config json.RawMessage) error {
// Check access to counter files
countersDir := filepath.Join(path, "counters")
- portCounterFiles := map[string]string{
- "ib_recv": filepath.Join(countersDir, "port_rcv_data"),
- "ib_xmit": filepath.Join(countersDir, "port_xmit_data"),
- "ib_recv_pkts": filepath.Join(countersDir, "port_rcv_packets"),
- "ib_xmit_pkts": filepath.Join(countersDir, "port_xmit_packets"),
+ portCounterFiles := map[string]InfinibandCollectorMetric{
+ "ib_recv": {path: filepath.Join(countersDir, "port_rcv_data"), unit: "bytes"},
+ "ib_xmit": {path: filepath.Join(countersDir, "port_xmit_data"), unit: "bytes"},
+ "ib_recv_pkts": {path: filepath.Join(countersDir, "port_rcv_packets"), unit: "packets"},
+ "ib_xmit_pkts": {path: filepath.Join(countersDir, "port_xmit_packets"), unit: "packets"},
}
- for _, counterFile := range portCounterFiles {
- err := unix.Access(counterFile, unix.R_OK)
+ for _, counter := range portCounterFiles {
+ err := unix.Access(counter.path, unix.R_OK)
if err != nil {
- return fmt.Errorf("unable to access %s: %v", counterFile, err)
+ return fmt.Errorf("unable to access %s: %v", counter.path, err)
}
}
@@ -165,14 +170,14 @@ func (m *InfinibandCollector) Read(interval time.Duration, output chan lp.CCMetr
m.lastTimestamp = now
for _, info := range m.info {
- for counterName, counterFile := range info.portCounterFiles {
+ for counterName, counterDef := range info.portCounterFiles {
// Read counter file
- line, err := ioutil.ReadFile(counterFile)
+ line, err := ioutil.ReadFile(counterDef.path)
if err != nil {
cclog.ComponentError(
m.name,
- fmt.Sprintf("Read(): Failed to read from file '%s': %v", counterFile, err))
+ fmt.Sprintf("Read(): Failed to read from file '%s': %v", counterDef.path, err))
continue
}
data := strings.TrimSpace(string(line))
@@ -189,6 +194,7 @@ func (m *InfinibandCollector) Read(interval time.Duration, output chan lp.CCMetr
// Send absolut values
if m.config.SendAbsoluteValues {
if y, err := lp.New(counterName, info.tagSet, m.meta, map[string]interface{}{"value": v}, now); err == nil {
+ y.AddMeta("unit", counterDef.unit)
output <- y
}
}
@@ -198,6 +204,7 @@ func (m *InfinibandCollector) Read(interval time.Duration, output chan lp.CCMetr
if info.lastState[counterName] >= 0 {
rate := float64((v - info.lastState[counterName])) / timeDiff
if y, err := lp.New(counterName+"_bw", info.tagSet, m.meta, map[string]interface{}{"value": rate}, now); err == nil {
+ y.AddMeta("unit", counterDef.unit+"/sec")
output <- y
}
}
diff --git a/collectors/ipmiMetric.go b/collectors/ipmiMetric.go
index e59f407..16b08ef 100644
--- a/collectors/ipmiMetric.go
+++ b/collectors/ipmiMetric.go
@@ -54,7 +54,7 @@ func (m *IpmiCollector) Init(config json.RawMessage) error {
m.ipmisensors = p
}
if len(m.ipmitool) == 0 && len(m.ipmisensors) == 0 {
- return errors.New("No IPMI reader found")
+ return errors.New("no IPMI reader found")
}
m.init = true
return nil
diff --git a/collectors/likwidMetric.go b/collectors/likwidMetric.go
index e0b0d7e..f2229d1 100644
--- a/collectors/likwidMetric.go
+++ b/collectors/likwidMetric.go
@@ -15,8 +15,12 @@ import (
"io/ioutil"
"math"
"os"
+ "os/signal"
+ "sort"
"strconv"
"strings"
+ "sync"
+ "syscall"
"time"
"unsafe"
@@ -46,6 +50,16 @@ type LikwidCollectorEventsetConfig struct {
Metrics []LikwidCollectorMetricConfig `json:"metrics"`
}
+type LikwidEventsetConfig struct {
+ internal int
+ gid C.int
+ eorder []*C.char
+ estr *C.char
+ go_estr string
+ results map[int]map[string]interface{}
+ metrics map[int]map[string]float64
+}
+
type LikwidCollectorConfig struct {
Eventsets []LikwidCollectorEventsetConfig `json:"eventsets"`
Metrics []LikwidCollectorMetricConfig `json:"globalmetrics,omitempty"`
@@ -58,17 +72,18 @@ type LikwidCollectorConfig struct {
type LikwidCollector struct {
metricCollector
- cpulist []C.int
- cpu2tid map[int]int
- sock2tid map[int]int
- metrics map[C.int]map[string]int
- groups []C.int
- config LikwidCollectorConfig
- results map[int]map[int]map[string]interface{}
- mresults map[int]map[int]map[string]float64
- gmresults map[int]map[string]float64
- basefreq float64
- running bool
+ cpulist []C.int
+ cpu2tid map[int]int
+ sock2tid map[int]int
+ metrics map[C.int]map[string]int
+ groups []C.int
+ config LikwidCollectorConfig
+ gmresults map[int]map[string]float64
+ basefreq float64
+ running bool
+ initialized bool
+ likwidGroups map[C.int]LikwidEventsetConfig
+ lock sync.Mutex
}
type LikwidMetric struct {
@@ -86,14 +101,60 @@ func eventsToEventStr(events map[string]string) string {
return strings.Join(elist, ",")
}
+func genLikwidEventSet(input LikwidCollectorEventsetConfig) LikwidEventsetConfig {
+ tmplist := make([]string, 0)
+ clist := make([]string, 0)
+ for k := range input.Events {
+ clist = append(clist, k)
+ }
+ sort.Strings(clist)
+ elist := make([]*C.char, 0)
+ for _, k := range clist {
+ v := input.Events[k]
+ tmplist = append(tmplist, fmt.Sprintf("%s:%s", v, k))
+ c_counter := C.CString(k)
+ elist = append(elist, c_counter)
+ }
+ estr := strings.Join(tmplist, ",")
+ res := make(map[int]map[string]interface{})
+ met := make(map[int]map[string]float64)
+ for _, i := range topo.CpuList() {
+ res[i] = make(map[string]interface{})
+ for k := range input.Events {
+ res[i][k] = 0.0
+ }
+ met[i] = make(map[string]float64)
+ for _, v := range input.Metrics {
+ res[i][v.Name] = 0.0
+ }
+ }
+ return LikwidEventsetConfig{
+ gid: -1,
+ eorder: elist,
+ estr: C.CString(estr),
+ go_estr: estr,
+ results: res,
+ metrics: met,
+ }
+}
+
+func testLikwidMetricFormula(formula string, params []string) bool {
+ myparams := make(map[string]interface{})
+ for _, p := range params {
+ myparams[p] = float64(1.0)
+ }
+ _, err := agg.EvalFloat64Condition(formula, myparams)
+ return err == nil
+}
+
func getBaseFreq() float64 {
+ files := []string{
+ "/sys/devices/system/cpu/cpu0/cpufreq/bios_limit",
+ "/sys/devices/system/cpu/cpu0/cpufreq/base_frequency",
+ }
var freq float64 = math.NaN()
- C.power_init(0)
- info := C.get_powerInfo()
- if float64(info.baseFrequency) != 0 {
- freq = float64(info.baseFrequency) * 1e6
- } else {
- buffer, err := ioutil.ReadFile("/sys/devices/system/cpu/cpu0/cpufreq/bios_limit")
+ for _, f := range files {
+ buffer, err := ioutil.ReadFile(f)
if err == nil {
data := strings.Replace(string(buffer), "\n", "", -1)
x, err := strconv.ParseInt(data, 0, 64)
@@ -102,12 +163,22 @@ func getBaseFreq() float64 {
}
}
}
+
+ if math.IsNaN(freq) {
+ C.power_init(0)
+ info := C.get_powerInfo()
+ if float64(info.baseFrequency) != 0 {
+ freq = float64(info.baseFrequency) * 1e6
+ }
+ C.power_finalize()
+ }
return freq
}
func (m *LikwidCollector) Init(config json.RawMessage) error {
- var ret C.int
m.name = "LikwidCollector"
+ m.initialized = false
+ m.running = false
m.config.AccessMode = LIKWID_DEF_ACCESSMODE
m.config.LibraryPath = LIKWID_LIB_NAME
if len(config) > 0 {
@@ -131,7 +202,7 @@ func (m *LikwidCollector) Init(config json.RawMessage) error {
}
m.setup()
- m.meta = map[string]string{"source": m.name, "group": "PerfCounter"}
+ m.meta = map[string]string{"group": "PerfCounter"}
cclog.ComponentDebug(m.name, "Get cpulist and init maps and lists")
cpulist := topo.CpuList()
m.cpulist = make([]C.int, len(cpulist))
@@ -139,164 +210,137 @@ func (m *LikwidCollector) Init(config json.RawMessage) error {
for i, c := range cpulist {
m.cpulist[i] = C.int(c)
m.cpu2tid[c] = i
-
}
- m.results = make(map[int]map[int]map[string]interface{})
- m.mresults = make(map[int]map[int]map[string]float64)
+
+ m.likwidGroups = make(map[C.int]LikwidEventsetConfig)
+
+ // m.results = make(map[int]map[int]map[string]interface{})
+ // m.mresults = make(map[int]map[int]map[string]float64)
m.gmresults = make(map[int]map[string]float64)
- cclog.ComponentDebug(m.name, "initialize LIKWID topology")
- ret = C.topology_init()
- if ret != 0 {
- err := errors.New("failed to initialize LIKWID topology")
- cclog.ComponentError(m.name, err.Error())
- return err
- }
-
- switch m.config.AccessMode {
- case "direct":
- C.HPMmode(0)
- case "accessdaemon":
- if len(m.config.DaemonPath) > 0 {
- p := os.Getenv("PATH")
- os.Setenv("PATH", m.config.DaemonPath+":"+p)
- }
- C.HPMmode(1)
- }
-
- cclog.ComponentDebug(m.name, "initialize LIKWID perfmon module")
- ret = C.perfmon_init(C.int(len(m.cpulist)), &m.cpulist[0])
- if ret != 0 {
- C.topology_finalize()
- err := errors.New("failed to initialize LIKWID topology")
- cclog.ComponentError(m.name, err.Error())
- return err
+ for _, tid := range m.cpu2tid {
+ m.gmresults[tid] = make(map[string]float64)
}
// This is for the global metrics computation test
- globalParams := make(map[string]interface{})
- globalParams["time"] = float64(1.0)
- globalParams["inverseClock"] = float64(1.0)
- // While adding the events, we test the metrics whether they can be computed at all
- for i, evset := range m.config.Eventsets {
- var gid C.int
- var cstr *C.char
+ totalMetrics := 0
+ // Generate parameter list for the metric computing test
+ params := make([]string, 0)
+ params = append(params, "time", "inverseClock")
+ // Generate parameter list for the global metric computing test
+ globalParams := make([]string, 0)
+ globalParams = append(globalParams, "time", "inverseClock")
+ // We test the eventset metrics whether they can be computed at all
+ for _, evset := range m.config.Eventsets {
if len(evset.Events) > 0 {
- estr := eventsToEventStr(evset.Events)
- // Generate parameter list for the metric computing test
- params := make(map[string]interface{})
- params["time"] = float64(1.0)
- params["inverseClock"] = float64(1.0)
+ params = params[:2]
for counter := range evset.Events {
- params[counter] = float64(1.0)
+ params = append(params, counter)
}
for _, metric := range evset.Metrics {
// Try to evaluate the metric
- _, err := agg.EvalFloat64Condition(metric.Calc, params)
- if err != nil {
- cclog.ComponentError(m.name, "Calculation for metric", metric.Name, "failed:", err.Error())
- continue
- }
- // If the metric is not in the parameter list for the global metrics, add it
- if _, ok := globalParams[metric.Name]; !ok {
- globalParams[metric.Name] = float64(1.0)
+ if testLikwidMetricFormula(metric.Calc, params) {
+ // Add the computable metric to the parameter list for the global metrics
+ globalParams = append(globalParams, metric.Name)
+ totalMetrics++
+ } else {
+ metric.Calc = ""
}
}
- // Now we add the list of events to likwid
- cstr = C.CString(estr)
- gid = C.perfmon_addEventSet(cstr)
} else {
cclog.ComponentError(m.name, "Invalid Likwid eventset config, no events given")
continue
}
- if gid >= 0 {
- m.groups = append(m.groups, gid)
- }
- C.free(unsafe.Pointer(cstr))
- m.results[i] = make(map[int]map[string]interface{})
- m.mresults[i] = make(map[int]map[string]float64)
- for tid := range m.cpulist {
- m.results[i][tid] = make(map[string]interface{})
- m.mresults[i][tid] = make(map[string]float64)
- if i == 0 {
- m.gmresults[tid] = make(map[string]float64)
- }
- }
}
for _, metric := range m.config.Metrics {
// Try to evaluate the global metric
- _, err := agg.EvalFloat64Condition(metric.Calc, globalParams)
- if err != nil {
- cclog.ComponentError(m.name, "Calculation for metric", metric.Name, "failed:", err.Error())
- continue
+ if !testLikwidMetricFormula(metric.Calc, globalParams) {
+ cclog.ComponentError(m.name, "Calculation for metric", metric.Name, "failed")
+ metric.Calc = ""
+ } else {
+ totalMetrics++
}
}
// If no event set could be added, shut down LikwidCollector
- if len(m.groups) == 0 {
- C.perfmon_finalize()
- C.topology_finalize()
- err := errors.New("no LIKWID performance group initialized")
+ if totalMetrics == 0 {
+ err := errors.New("no LIKWID eventset or metric usable")
cclog.ComponentError(m.name, err.Error())
return err
}
- m.basefreq = getBaseFreq()
- cclog.ComponentDebug(m.name, "BaseFreq", m.basefreq)
m.init = true
return nil
}
// take a measurement for 'interval' seconds of event set index 'group'
-func (m *LikwidCollector) takeMeasurement(group int, interval time.Duration) error {
+func (m *LikwidCollector) takeMeasurement(evset LikwidEventsetConfig, interval time.Duration) (bool, error) {
var ret C.int
- gid := m.groups[group]
- ret = C.perfmon_setupCounters(gid)
- if ret != 0 {
- gctr := C.GoString(C.perfmon_getGroupName(gid))
- err := fmt.Errorf("failed to setup performance group %d (%s)", gid, gctr)
- return err
+ m.lock.Lock()
+ if m.initialized {
+ ret = C.perfmon_setupCounters(evset.gid)
+ if ret != 0 {
+ var err error = nil
+ var skip bool = false
+ if ret == -37 {
+ skip = true
+ } else {
+ err = fmt.Errorf("failed to setup performance group %d", evset.gid)
+ }
+ m.lock.Unlock()
+ return skip, err
+ }
+ ret = C.perfmon_startCounters()
+ if ret != 0 {
+ var err error = nil
+ var skip bool = false
+ if ret == -37 {
+ skip = true
+ } else {
+ err = fmt.Errorf("failed to setup performance group %d", evset.gid)
+ }
+ m.lock.Unlock()
+ return skip, err
+ }
+ m.running = true
+ time.Sleep(interval)
+ m.running = false
+ ret = C.perfmon_stopCounters()
+ if ret != 0 {
+ var err error = nil
+ var skip bool = false
+ if ret == -37 {
+ skip = true
+ } else {
+ err = fmt.Errorf("failed to setup performance group %d", evset.gid)
+ }
+ m.lock.Unlock()
+ return skip, err
+ }
}
- ret = C.perfmon_startCounters()
- if ret != 0 {
- gctr := C.GoString(C.perfmon_getGroupName(gid))
- err := fmt.Errorf("failed to start performance group %d (%s)", gid, gctr)
- return err
- }
- m.running = true
- time.Sleep(interval)
- m.running = false
- ret = C.perfmon_stopCounters()
- if ret != 0 {
- gctr := C.GoString(C.perfmon_getGroupName(gid))
- err := fmt.Errorf("failed to stop performance group %d (%s)", gid, gctr)
- return err
- }
- return nil
+ m.lock.Unlock()
+ return false, nil
}
// Get all measurement results for an event set, derive the metric values out of the measurement results and send it
-func (m *LikwidCollector) calcEventsetMetrics(group int, interval time.Duration, output chan lp.CCMetric) error {
- var eidx C.int
- evset := m.config.Eventsets[group]
- gid := m.groups[group]
+func (m *LikwidCollector) calcEventsetMetrics(evset LikwidEventsetConfig, interval time.Duration, output chan lp.CCMetric) error {
invClock := float64(1.0 / m.basefreq)
// Go over events and get the results
- for eidx = 0; int(eidx) < len(evset.Events); eidx++ {
- ctr := C.perfmon_getCounterName(gid, eidx)
- gctr := C.GoString(ctr)
-
+ for eidx, counter := range evset.eorder {
+ gctr := C.GoString(counter)
for _, tid := range m.cpu2tid {
- if tid >= 0 {
- m.results[group][tid]["time"] = interval.Seconds()
- m.results[group][tid]["inverseClock"] = invClock
- res := C.perfmon_getLastResult(gid, eidx, C.int(tid))
- m.results[group][tid][gctr] = float64(res)
+ res := C.perfmon_getLastResult(evset.gid, C.int(eidx), C.int(tid))
+ fres := float64(res)
+ if m.config.InvalidToZero && (math.IsNaN(fres) || math.IsInf(fres, 0)) {
+ fres = 0.0
}
+ evset.results[tid][gctr] = fres
+ evset.results[tid]["time"] = interval.Seconds()
+ evset.results[tid]["inverseClock"] = invClock
}
}
// Go over the event set metrics, derive the value out of the event:counter values and send it
- for _, metric := range evset.Metrics {
+ for _, metric := range m.config.Eventsets[evset.internal].Metrics {
// The metric scope is determined in the Init() function
// Get the map scope-id -> tids
scopemap := m.cpu2tid
@@ -304,19 +348,16 @@ func (m *LikwidCollector) calcEventsetMetrics(group int, interval time.Duration,
scopemap = m.sock2tid
}
for domain, tid := range scopemap {
- if tid >= 0 {
- value, err := agg.EvalFloat64Condition(metric.Calc, m.results[group][tid])
+ if tid >= 0 && len(metric.Calc) > 0 {
+ value, err := agg.EvalFloat64Condition(metric.Calc, evset.results[tid])
if err != nil {
cclog.ComponentError(m.name, "Calculation for metric", metric.Name, "failed:", err.Error())
- continue
- }
- m.mresults[group][tid][metric.Name] = value
- if m.config.InvalidToZero && math.IsNaN(value) {
value = 0.0
}
- if m.config.InvalidToZero && math.IsInf(value, 0) {
+ if m.config.InvalidToZero && (math.IsNaN(value) || math.IsInf(value, 0)) {
value = 0.0
}
+ evset.metrics[tid][metric.Name] = value
// Now we have the result, send it with the proper tags
if !math.IsNaN(value) {
if metric.Publish {
@@ -351,8 +392,8 @@ func (m *LikwidCollector) calcGlobalMetrics(interval time.Duration, output chan
if tid >= 0 {
// Here we generate parameter list
params := make(map[string]interface{})
- for j := range m.groups {
- for mname, mres := range m.mresults[j][tid] {
+ for _, evset := range m.likwidGroups {
+ for mname, mres := range evset.metrics[tid] {
params[mname] = mres
}
}
@@ -360,15 +401,12 @@ func (m *LikwidCollector) calcGlobalMetrics(interval time.Duration, output chan
value, err := agg.EvalFloat64Condition(metric.Calc, params)
if err != nil {
cclog.ComponentError(m.name, "Calculation for metric", metric.Name, "failed:", err.Error())
- continue
+ value = 0.0
+ }
+ if m.config.InvalidToZero && (math.IsNaN(value) || math.IsInf(value, 0)) {
+ value = 0.0
}
m.gmresults[tid][metric.Name] = value
- if m.config.InvalidToZero && math.IsNaN(value) {
- value = 0.0
- }
- if m.config.InvalidToZero && math.IsInf(value, 0) {
- value = 0.0
- }
// Now we have the result, send it with the proper tags
if !math.IsNaN(value) {
if metric.Publish {
@@ -392,38 +430,163 @@ func (m *LikwidCollector) calcGlobalMetrics(interval time.Duration, output chan
return nil
}
+func (m *LikwidCollector) LateInit() error {
+ var ret C.int
+ if m.initialized {
+ return nil
+ }
+ switch m.config.AccessMode {
+ case "direct":
+ C.HPMmode(0)
+ case "accessdaemon":
+ if len(m.config.DaemonPath) > 0 {
+ p := os.Getenv("PATH")
+ os.Setenv("PATH", m.config.DaemonPath+":"+p)
+ }
+ C.HPMmode(1)
+ }
+ cclog.ComponentDebug(m.name, "initialize LIKWID topology")
+ ret = C.topology_init()
+ if ret != 0 {
+ err := errors.New("failed to initialize LIKWID topology")
+ cclog.ComponentError(m.name, err.Error())
+ return err
+ }
+
+ m.sock2tid = make(map[int]int)
+ tmp := make([]C.int, 1)
+ for _, sid := range topo.SocketList() {
+ cstr := C.CString(fmt.Sprintf("S%d:0", sid))
+ ret = C.cpustr_to_cpulist(cstr, &tmp[0], 1)
+ if ret > 0 {
+ m.sock2tid[sid] = m.cpu2tid[int(tmp[0])]
+ }
+ C.free(unsafe.Pointer(cstr))
+ }
+
+ m.basefreq = getBaseFreq()
+ cclog.ComponentDebug(m.name, "BaseFreq", m.basefreq)
+
+ cclog.ComponentDebug(m.name, "initialize LIKWID perfmon module")
+ ret = C.perfmon_init(C.int(len(m.cpulist)), &m.cpulist[0])
+ if ret != 0 {
+ var err error = nil
+ C.topology_finalize()
+ if ret != -22 {
+ err = errors.New("failed to initialize LIKWID perfmon")
+ cclog.ComponentError(m.name, err.Error())
+ } else {
+ err = errors.New("access to LIKWID perfmon locked")
+ }
+ return err
+ }
+
+ // While adding the events, we test the metrics whether they can be computed at all
+ for i, evset := range m.config.Eventsets {
+ var gid C.int
+ if len(evset.Events) > 0 {
+ skip := false
+ likwidGroup := genLikwidEventSet(evset)
+ for _, g := range m.likwidGroups {
+ if likwidGroup.go_estr == g.go_estr {
+ skip = true
+ break
+ }
+ }
+ if skip {
+ continue
+ }
+ // Now we add the list of events to likwid
+ gid = C.perfmon_addEventSet(likwidGroup.estr)
+ if gid >= 0 {
+ likwidGroup.gid = gid
+ likwidGroup.internal = i
+ m.likwidGroups[gid] = likwidGroup
+ }
+ } else {
+ cclog.ComponentError(m.name, "Invalid Likwid eventset config, no events given")
+ continue
+ }
+
+ }
+
+ // If no event set could be added, shut down LikwidCollector
+ if len(m.likwidGroups) == 0 {
+ C.perfmon_finalize()
+ C.topology_finalize()
+ err := errors.New("no LIKWID performance group initialized")
+ cclog.ComponentError(m.name, err.Error())
+ return err
+ }
+ sigchan := make(chan os.Signal, 1)
+ signal.Notify(sigchan, syscall.SIGCHLD)
+ signal.Notify(sigchan, os.Interrupt)
+ go func() {
+ <-sigchan
+
+ signal.Stop(sigchan)
+ m.initialized = false
+ }()
+ m.initialized = true
+ return nil
+}
+
// main read function taking multiple measurement rounds, each 'interval' seconds long
func (m *LikwidCollector) Read(interval time.Duration, output chan lp.CCMetric) {
+ var skip bool = false
+ var err error
if !m.init {
return
}
- for i := range m.groups {
- // measure event set 'i' for 'interval' seconds
- err := m.takeMeasurement(i, interval)
+ if !m.initialized {
+ m.lock.Lock()
+ err = m.LateInit()
if err != nil {
- cclog.ComponentError(m.name, err.Error())
+ m.lock.Unlock()
return
}
- // read measurements and derive event set metrics
- m.calcEventsetMetrics(i, interval, output)
+ m.initialized = true
+ m.lock.Unlock()
+ }
+
+ if m.initialized && !skip {
+ for _, evset := range m.likwidGroups {
+ if !skip {
+ // measure event set 'i' for 'interval' seconds
+ skip, err = m.takeMeasurement(evset, interval)
+ if err != nil {
+ cclog.ComponentError(m.name, err.Error())
+ return
+ }
+ }
+
+ if !skip {
+ // read measurements and derive event set metrics
+ m.calcEventsetMetrics(evset, interval, output)
+ }
+ }
+ if !skip {
+ // use the event set metrics to derive the global metrics
+ m.calcGlobalMetrics(interval, output)
+ }
}
- // use the event set metrics to derive the global metrics
- m.calcGlobalMetrics(interval, output)
}
func (m *LikwidCollector) Close() {
if m.init {
- cclog.ComponentDebug(m.name, "Closing ...")
m.init = false
- if m.running {
- cclog.ComponentDebug(m.name, "Stopping counters")
- C.perfmon_stopCounters()
+ cclog.ComponentDebug(m.name, "Closing ...")
+ m.lock.Lock()
+ if m.initialized {
+ cclog.ComponentDebug(m.name, "Finalize LIKWID perfmon module")
+ C.perfmon_finalize()
+ m.initialized = false
}
- cclog.ComponentDebug(m.name, "Finalize LIKWID perfmon module")
- C.perfmon_finalize()
+ m.lock.Unlock()
cclog.ComponentDebug(m.name, "Finalize LIKWID topology module")
C.topology_finalize()
+
cclog.ComponentDebug(m.name, "Closing done")
}
}
diff --git a/collectors/likwidMetric.md b/collectors/likwidMetric.md
index fe28857..2d622d1 100644
--- a/collectors/likwidMetric.md
+++ b/collectors/likwidMetric.md
@@ -3,32 +3,63 @@
The `likwid` collector is probably the most complicated collector. The LIKWID library is included as static library with *direct* access mode. The *direct* access mode is suitable if the daemon is executed by a root user. The static library does not contain the performance groups, so all information needs to be provided in the configuration.
-The `likwid` configuration consists of two parts, the "eventsets" and "globalmetrics":
-- An event set list itself has two parts, the "events" and a set of derivable "metrics". Each of the "events" is a counter:event pair in LIKWID's syntax. The "metrics" are a list of formulas to derive the metric value from the measurements of the "events". Each metric has a name, the formula, a scope and a publish flag. Counter names can be used like variables in the formulas, so `PMC0+PMC1` sums the measurements for the both events configured in the counters `PMC0` and `PMC1`. The scope tells the Collector whether it is a metric for each hardware thread (`cpu`) or each CPU socket (`socket`). You may specify a unit for the metric with `unit`. The last one is the publishing flag. It tells the collector whether a metric should be sent to the router.
-- The global metrics are metrics which require data from all event set measurements to be derived. The inputs are the metrics in the event sets. Similar to the metrics in the event sets, the global metrics are defined by a name, a formula, a scope and a publish flag. See event set metrics for details. The only difference is that there is no access to the raw event measurements anymore but only to the metrics. So, the idea is to derive a metric in the "eventsets" section and reuse it in the "globalmetrics" part. If you need a metric only for deriving the global metrics, disable forwarding of the event set metrics (`publish=false`). **Be aware** that the combination might be misleading because the "behavior" of a metric changes over time and the multiple measurements might count different computing phases. Similar to the metrics in the eventset, you can specify a metric unit with the `unit` field.
+```json
+ "likwid": {
+ "force_overwrite" : false,
+ "invalid_to_zero" : false,
+ "eventsets": [
+ {
+ "events" : {
+ "COUNTER0": "EVENT0",
+ "COUNTER1": "EVENT1",
+ },
+ "metrics" : [
+ {
+ "name": "sum_01",
+ "calc": "COUNTER0 + COUNTER1",
+ "publish": false,
+ "unit": "myunit",
+ "type": "cpu"
+ }
+ ]
+ }
+ ]
+ "globalmetrics" : [
+ {
+ "name": "global_sum",
+ "calc": "sum_01",
+ "publish": true,
+ "unit": "myunit",
+ "type": "cpu"
+ }
+ ]
+ }
+```
+
+The `likwid` configuration consists of two parts, the `eventsets` and `globalmetrics`:
+- An event set list itself has two parts, the `events` and a set of derivable `metrics`. Each of the `events` is a `counter:event` pair in LIKWID's syntax. The `metrics` are a list of formulas to derive the metric value from the measurements of the `events`' values. Each metric has a name, the formula, a type and a publish flag. There is an optional `unit` field. Counter names can be used like variables in the formulas, so `PMC0+PMC1` sums the measurements for the both events configured in the counters `PMC0` and `PMC1`. You can optionally use `time` for the measurement time and `inverseClock` for `1.0/baseCpuFrequency`. The type tells the LikwidCollector whether it is a metric for each hardware thread (`cpu`) or each CPU socket (`socket`). You may specify a unit for the metric with `unit`. The last one is the publishing flag. It tells the LikwidCollector whether a metric should be sent to the router or is only used internally to compute a global metric.
+- The `globalmetrics` are metrics which require data from multiple event set measurements to be derived. The inputs are the metrics in the event sets. Similar to the metrics in the event sets, the global metrics are defined by a name, a formula, a scope and a publish flag. See event set metrics for details. The only difference is that there is no access to the raw event measurements anymore but only to the metrics. Also `time` and `inverseClock` cannot be used anymore. So, the idea is to derive a metric in the `eventsets` section and reuse it in the `globalmetrics` part. If you need a metric only for deriving the global metrics, disable forwarding of the event set metrics (`"publish": false`). **Be aware** that the combination might be misleading because the "behavior" of a metric changes over time and the multiple measurements might count different computing phases. Similar to the metrics in the eventset, you can specify a metric unit with the `unit` field.
Additional options:
-- `access_mode` : Method to use for hardware performance monitoring (`direct` access as root user, `accessdaemon` for the daemon mode)
-- `accessdaemon_path`: Folder with the access daemon `likwid-accessD`, commonly `$LIKWID_INSTALL_LOC/sbin`
- `force_overwrite`: Same as setting `LIKWID_FORCE=1`. In case counters are already in-use, LIKWID overwrites their configuration to do its measurements
-- `invalid_to_zero`: In some cases, the calculations result in `NaN` or `Inf`. With this option, all `NaN` and `Inf` values are replaces with `0.0`.
-- `access_mode`: Specify LIKWID access mode: `direct` for direct register access as root user or `accessdaemon`
-- `accessdaemon_path`: Folder of the accessDaemon `likwid-accessD`
-- `liblikwid_path`: Location of `liblikwid.so`
+- `invalid_to_zero`: In some cases, the calculations result in `NaN` or `Inf`. With this option, all `NaN` and `Inf` values are replaces with `0.0`. See below in [seperate section](./likwidMetric.md#invalid_to_zero-option)
+- `access_mode`: Specify LIKWID access mode: `direct` for direct register access as root user or `accessdaemon`. The access mode `perf_event` is current untested.
+- `accessdaemon_path`: Folder of the accessDaemon `likwid-accessD` (like `/usr/local/sbin`)
+- `liblikwid_path`: Location of `liblikwid.so` including file name like `/usr/local/lib/liblikwid.so`
### Available metric scopes
-Hardware performance counters are scattered all over the system nowadays. A counter coveres a specific part of the system. While there are hardware thread specific counter for CPU cycles, instructions and so on, some others are specific for a whole CPU socket/package. To address that, the collector provides the specification of a 'scope' for each metric.
+Hardware performance counters are scattered all over the system nowadays. A counter coveres a specific part of the system. While there are hardware thread specific counter for CPU cycles, instructions and so on, some others are specific for a whole CPU socket/package. To address that, the LikwidCollector provides the specification of a `type` for each metric.
- `cpu` : One metric per CPU hardware thread with the tags `"type" : "cpu"` and `"type-id" : "$cpu_id"`
- `socket` : One metric per CPU socket/package with the tags `"type" : "socket"` and `"type-id" : "$socket_id"`
-**Note:** You cannot specify `socket` scope for a metric that is measured at `cpu` scope, so some kind of expert knowledge or lookup work in the [Likwid Wiki](https://github.com/RRZE-HPC/likwid/wiki) is required. Get the scope of each counter from the *Architecture* pages and as soon as one counter in a metric is socket-specific, the whole metric is socket-specific.
+**Note:** You should not specify the `socket` type for a metric that is measured at `cpu` scope and vice versa, so some kind of expert knowledge or lookup work in the [Likwid Wiki](https://github.com/RRZE-HPC/likwid/wiki) is required. Get the scope of each counter from the *Architecture* pages and as soon as one counter in a metric is socket-specific, the whole metric is socket-specific.
As a guideline:
- All counters `FIXCx`, `PMCy` and `TMAz` have the scope `cpu`
- All counters names containing `BOX` have the scope `socket`
-- All `PWRx` counters have scope `socket`, except `"PWR1" : "RAPL_CORE_ENERGY"` has `cpu` scope
+- All `PWRx` counters have scope `socket`, except `"PWR1" : "RAPL_CORE_ENERGY"` has `cpu` scope (AMD Zen)
- All `DFCx` counters have scope `socket`
### Help with the configuration
@@ -50,6 +81,7 @@ $ scripts/likwid_perfgroup_to_cc_config.py ICX MEM_DP
{
"events": {
"FIXC0": "INSTR_RETIRED_ANY",
+ "FIXC1": "CPU_CLK_UNHALTED_CORE",
"..." : "..."
},
"metrics" : [
@@ -75,21 +107,28 @@ LIKWID checks the file `/var/run/likwid.lock` before performing any interfering
Before (SLURM prolog, ...)
```
-$ chwon $JOBUSER /var/run/likwid.lock
+$ chown $JOBUSER /var/run/likwid.lock
```
After (SLURM epilog, ...)
```
-$ chwon $CCUSER /var/run/likwid.lock
+$ chown $CCUSER /var/run/likwid.lock
```
+### `invalid_to_zero` option
+In some cases LIKWID returns `0.0` for some events that are further used in processing and maybe used as divisor in a calculation. After evaluation of a metric, the result might be `NaN` or `+-Inf`. These resulting metrics are commonly not created and forwarded to the router because the [InfluxDB line protocol](https://docs.influxdata.com/influxdb/cloud/reference/syntax/line-protocol/#float) does not support these special floating-point values. If you want to have them sent, this option forces these metric values to be `0.0` instead.
+
+One might think this does not happen often but often used metrics in the world of performance engineering like Instructions-per-Cycle (IPC) or more frequently the actual CPU clock are derived with events like `CPU_CLK_UNHALTED_CORE` (Intel) which do not increment in halted state (as the name implies). In there are different power management systems in a chip which can cause a hardware thread to go in such a state. Moreover, if no cycles are executed by the core, also many other events are not incremented as well (like `INSTR_RETIRED_ANY` for retired instructions and part of IPC).
+
+
### Example configuration
+#### AMD Zen3
```json
"likwid": {
"force_overwrite" : false,
- "nan_to_zero" : false,
+ "invalid_to_zero" : false,
"eventsets": [
{
"events": {
@@ -180,33 +219,3 @@ $ chwon $CCUSER /var/run/likwid.lock
}
```
-### How to get the eventsets and metrics from LIKWID
-
-The `likwid` collector reads hardware performance counters at a **cpu** and **socket** level. The configuration looks quite complicated but it is basically copy&paste from [LIKWID's performance groups](https://github.com/RRZE-HPC/likwid/tree/master/groups). The collector made multiple iterations and tried to use the performance groups but it lacked flexibility. The current way of configuration provides most flexibility.
-
-The logic is as following: There are multiple eventsets, each consisting of a list of counters+events and a list of metrics. If you compare a common performance group with the example setting above, there is not much difference:
-```
-EVENTSET -> "events": {
-FIXC1 ACTUAL_CPU_CLOCK -> "FIXC1": "ACTUAL_CPU_CLOCK",
-FIXC2 MAX_CPU_CLOCK -> "FIXC2": "MAX_CPU_CLOCK",
-PMC0 RETIRED_INSTRUCTIONS -> "PMC0" : "RETIRED_INSTRUCTIONS",
-PMC1 CPU_CLOCKS_UNHALTED -> "PMC1" : "CPU_CLOCKS_UNHALTED",
-PMC2 RETIRED_SSE_AVX_FLOPS_ALL -> "PMC2": "RETIRED_SSE_AVX_FLOPS_ALL",
-PMC3 MERGE -> "PMC3": "MERGE",
- -> }
-```
-
-The metrics are following the same procedure:
-
-```
-METRICS -> "metrics": [
-IPC PMC0/PMC1 -> {
- -> "name" : "IPC",
- -> "calc" : "PMC0/PMC1",
- -> "scope": "cpu",
- -> "publish": true
- -> }
- -> ]
-```
-
-The script `scripts/likwid_perfgroup_to_cc_config.py` might help you.
\ No newline at end of file
diff --git a/collectors/lustreMetric.go b/collectors/lustreMetric.go
index 67efd7a..d5a96e4 100644
--- a/collectors/lustreMetric.go
+++ b/collectors/lustreMetric.go
@@ -19,23 +19,31 @@ const LCTL_CMD = `lctl`
const LCTL_OPTION = `get_param`
type LustreCollectorConfig struct {
- LCtlCommand string `json:"lctl_command"`
- ExcludeMetrics []string `json:"exclude_metrics"`
- SendAllMetrics bool `json:"send_all_metrics"`
- Sudo bool `json:"use_sudo"`
- SendAbsoluteValues bool `json:"send_abs_values"`
- SendDerivedValues bool `json:"send_derived_values"`
+ LCtlCommand string `json:"lctl_command,omitempty"`
+ ExcludeMetrics []string `json:"exclude_metrics,omitempty"`
+ Sudo bool `json:"use_sudo,omitempty"`
+ SendAbsoluteValues bool `json:"send_abs_values,omitempty"`
+ SendDerivedValues bool `json:"send_derived_values,omitempty"`
+ SendDiffValues bool `json:"send_diff_values,omitempty"`
+}
+
+type LustreMetricDefinition struct {
+ name string
+ lineprefix string
+ lineoffset int
+ unit string
+ calc string
}
type LustreCollector struct {
metricCollector
tags map[string]string
- matches map[string]map[string]int
- stats map[string]map[string]int64
config LustreCollectorConfig
lctl string
sudoCmd string
- lastTimestamp time.Time // Store time stamp of last tick to derive bandwidths
+ lastTimestamp time.Time // Store time stamp of last tick to derive bandwidths
+ definitions []LustreMetricDefinition // Combined list without excluded metrics
+ stats map[string]map[string]int64 // Data for last value per device and metric
}
func (m *LustreCollector) getDeviceDataCommand(device string) []string {
@@ -78,6 +86,16 @@ func (m *LustreCollector) getDevices() []string {
return devices
}
+func getMetricData(lines []string, prefix string, offset int) (int64, error) {
+ for _, line := range lines {
+ if strings.HasPrefix(line, prefix) {
+ lf := strings.Fields(line)
+ return strconv.ParseInt(lf[offset], 0, 64)
+ }
+ }
+ return 0, errors.New("no such line in data")
+}
+
// //Version reading the stats data of a device from sysfs
// func (m *LustreCollector) getDeviceDataSysfs(device string) []string {
// llitedir := filepath.Join(LUSTRE_SYSFS, "llite")
@@ -90,6 +108,183 @@ func (m *LustreCollector) getDevices() []string {
// return strings.Split(string(buffer), "\n")
// }
+var LustreAbsMetrics = []LustreMetricDefinition{
+ {
+ name: "lustre_read_requests",
+ lineprefix: "read_bytes",
+ lineoffset: 1,
+ unit: "requests",
+ calc: "none",
+ },
+ {
+ name: "lustre_write_requests",
+ lineprefix: "write_bytes",
+ lineoffset: 1,
+ unit: "requests",
+ calc: "none",
+ },
+ {
+ name: "lustre_read_bytes",
+ lineprefix: "read_bytes",
+ lineoffset: 6,
+ unit: "bytes",
+ calc: "none",
+ },
+ {
+ name: "lustre_write_bytes",
+ lineprefix: "write_bytes",
+ lineoffset: 6,
+ unit: "bytes",
+ calc: "none",
+ },
+ {
+ name: "lustre_open",
+ lineprefix: "open",
+ lineoffset: 1,
+ unit: "",
+ calc: "none",
+ },
+ {
+ name: "lustre_close",
+ lineprefix: "close",
+ lineoffset: 1,
+ unit: "",
+ calc: "none",
+ },
+ {
+ name: "lustre_setattr",
+ lineprefix: "setattr",
+ lineoffset: 1,
+ unit: "",
+ calc: "none",
+ },
+ {
+ name: "lustre_getattr",
+ lineprefix: "getattr",
+ lineoffset: 1,
+ unit: "",
+ calc: "none",
+ },
+ {
+ name: "lustre_statfs",
+ lineprefix: "statfs",
+ lineoffset: 1,
+ unit: "",
+ calc: "none",
+ },
+ {
+ name: "lustre_inode_permission",
+ lineprefix: "inode_permission",
+ lineoffset: 1,
+ unit: "",
+ calc: "none",
+ },
+}
+
+var LustreDiffMetrics = []LustreMetricDefinition{
+ {
+ name: "lustre_read_requests_diff",
+ lineprefix: "read_bytes",
+ lineoffset: 1,
+ unit: "requests",
+ calc: "difference",
+ },
+ {
+ name: "lustre_write_requests_diff",
+ lineprefix: "write_bytes",
+ lineoffset: 1,
+ unit: "requests",
+ calc: "difference",
+ },
+ {
+ name: "lustre_read_bytes_diff",
+ lineprefix: "read_bytes",
+ lineoffset: 6,
+ unit: "bytes",
+ calc: "difference",
+ },
+ {
+ name: "lustre_write_bytes_diff",
+ lineprefix: "write_bytes",
+ lineoffset: 6,
+ unit: "bytes",
+ calc: "difference",
+ },
+ {
+ name: "lustre_open_diff",
+ lineprefix: "open",
+ lineoffset: 1,
+ unit: "",
+ calc: "difference",
+ },
+ {
+ name: "lustre_close_diff",
+ lineprefix: "close",
+ lineoffset: 1,
+ unit: "",
+ calc: "difference",
+ },
+ {
+ name: "lustre_setattr_diff",
+ lineprefix: "setattr",
+ lineoffset: 1,
+ unit: "",
+ calc: "difference",
+ },
+ {
+ name: "lustre_getattr_diff",
+ lineprefix: "getattr",
+ lineoffset: 1,
+ unit: "",
+ calc: "difference",
+ },
+ {
+ name: "lustre_statfs_diff",
+ lineprefix: "statfs",
+ lineoffset: 1,
+ unit: "",
+ calc: "difference",
+ },
+ {
+ name: "lustre_inode_permission_diff",
+ lineprefix: "inode_permission",
+ lineoffset: 1,
+ unit: "",
+ calc: "difference",
+ },
+}
+
+var LustreDeriveMetrics = []LustreMetricDefinition{
+ {
+ name: "lustre_read_requests_rate",
+ lineprefix: "read_bytes",
+ lineoffset: 1,
+ unit: "requests/sec",
+ calc: "derivative",
+ },
+ {
+ name: "lustre_write_requests_rate",
+ lineprefix: "write_bytes",
+ lineoffset: 1,
+ unit: "requests/sec",
+ calc: "derivative",
+ },
+ {
+ name: "lustre_read_bw",
+ lineprefix: "read_bytes",
+ lineoffset: 6,
+ unit: "bytes/sec",
+ calc: "derivative",
+ },
+ {
+ name: "lustre_write_bw",
+ lineprefix: "write_bytes",
+ lineoffset: 6,
+ unit: "bytes/sec",
+ calc: "derivative",
+ },
+}
+
func (m *LustreCollector) Init(config json.RawMessage) error {
var err error
m.name = "LustreCollector"
@@ -102,17 +297,9 @@ func (m *LustreCollector) Init(config json.RawMessage) error {
m.setup()
m.tags = map[string]string{"type": "node"}
m.meta = map[string]string{"source": m.name, "group": "Lustre"}
- defmatches := map[string]map[string]int{
- "read_bytes": {"lustre_read_bytes": 6, "lustre_read_requests": 1},
- "write_bytes": {"lustre_write_bytes": 6, "lustre_write_requests": 1},
- "open": {"lustre_open": 1},
- "close": {"lustre_close": 1},
- "setattr": {"lustre_setattr": 1},
- "getattr": {"lustre_getattr": 1},
- "statfs": {"lustre_statfs": 1},
- "inode_permission": {"lustre_inode_permission": 1}}
// Lustre file system statistics can only be queried by user root
+ // or with password-less sudo
if !m.config.Sudo {
user, err := user.Current()
if err != nil {
@@ -123,23 +310,15 @@ func (m *LustreCollector) Init(config json.RawMessage) error {
cclog.ComponentError(m.name, "Lustre file system statistics can only be queried by user root")
return err
}
+ } else {
+ p, err := exec.LookPath("sudo")
+ if err != nil {
+ cclog.ComponentError(m.name, "Cannot find 'sudo'")
+ return err
+ }
+ m.sudoCmd = p
}
- m.matches = make(map[string]map[string]int)
- for lineprefix, names := range defmatches {
- for metricname, offset := range names {
- _, skip := stringArrayContains(m.config.ExcludeMetrics, metricname)
- if skip {
- continue
- }
- if _, prefixExist := m.matches[lineprefix]; !prefixExist {
- m.matches[lineprefix] = make(map[string]int)
- }
- if _, metricExist := m.matches[lineprefix][metricname]; !metricExist {
- m.matches[lineprefix][metricname] = offset
- }
- }
- }
p, err := exec.LookPath(m.config.LCtlCommand)
if err != nil {
p, err = exec.LookPath(LCTL_CMD)
@@ -148,23 +327,47 @@ func (m *LustreCollector) Init(config json.RawMessage) error {
}
}
m.lctl = p
- if m.config.Sudo {
- p, err := exec.LookPath("sudo")
- if err != nil {
- m.sudoCmd = p
+
+ m.definitions = []LustreMetricDefinition{}
+ if m.config.SendAbsoluteValues {
+ for _, def := range LustreAbsMetrics {
+ if _, skip := stringArrayContains(m.config.ExcludeMetrics, def.name); !skip {
+ m.definitions = append(m.definitions, def)
+ }
}
}
+ if m.config.SendDiffValues {
+ for _, def := range LustreDiffMetrics {
+ if _, skip := stringArrayContains(m.config.ExcludeMetrics, def.name); !skip {
+ m.definitions = append(m.definitions, def)
+ }
+ }
+ }
+ if m.config.SendDerivedValues {
+ for _, def := range LustreDeriveMetrics {
+ if _, skip := stringArrayContains(m.config.ExcludeMetrics, def.name); !skip {
+ m.definitions = append(m.definitions, def)
+ }
+ }
+ }
+ if len(m.definitions) == 0 {
+ return errors.New("no metrics to collect")
+ }
devices := m.getDevices()
if len(devices) == 0 {
- return errors.New("no metrics to collect")
+ return errors.New("no Lustre devices found")
}
m.stats = make(map[string]map[string]int64)
for _, d := range devices {
m.stats[d] = make(map[string]int64)
- for _, names := range m.matches {
- for metricname := range names {
- m.stats[d][metricname] = 0
+ data := m.getDeviceDataCommand(d)
+ for _, def := range m.definitions {
+ x, err := getMetricData(data, def.lineprefix, def.lineoffset)
+ if err == nil {
+ m.stats[d][def.name] = x
+ } else {
+ m.stats[d][def.name] = 0
}
}
}
@@ -180,63 +383,43 @@ func (m *LustreCollector) Read(interval time.Duration, output chan lp.CCMetric)
now := time.Now()
tdiff := now.Sub(m.lastTimestamp)
for device, devData := range m.stats {
- stats := m.getDeviceDataCommand(device)
- processed := []string{}
-
- for _, line := range stats {
- lf := strings.Fields(line)
- if len(lf) > 1 {
- if fields, ok := m.matches[lf[0]]; ok {
- for name, idx := range fields {
- x, err := strconv.ParseInt(lf[idx], 0, 64)
- if err == nil {
- value := x - devData[name]
- devData[name] = x
- if value < 0 {
- value = 0
- }
- if m.config.SendAbsoluteValues {
- y, err := lp.New(name, m.tags, m.meta, map[string]interface{}{"value": value}, time.Now())
- if err == nil {
- y.AddTag("device", device)
- if strings.Contains(name, "byte") {
- y.AddMeta("unit", "Byte")
- }
- output <- y
- if m.config.SendAllMetrics {
- processed = append(processed, name)
- }
- }
- }
- if m.config.SendDerivedValues && strings.Contains(name, "bytes") {
- y, err := lp.New(name+"_bw", m.tags, m.meta, map[string]interface{}{"value": float64(value) / tdiff.Seconds()}, time.Now())
- if err == nil {
- y.AddTag("device", device)
- y.AddMeta("unit", "Bytes/sec")
- output <- y
- if m.config.SendAllMetrics {
- processed = append(processed, name)
- }
- }
- }
- }
- }
- }
+ data := m.getDeviceDataCommand(device)
+ for _, def := range m.definitions {
+ var use_x int64
+ var err error
+ var y lp.CCMetric
+ x, err := getMetricData(data, def.lineprefix, def.lineoffset)
+ if err == nil {
+ use_x = x
+ } else {
+ use_x = devData[def.name]
}
- }
- if m.config.SendAllMetrics {
- for name := range devData {
- if _, done := stringArrayContains(processed, name); !done {
- y, err := lp.New(name, m.tags, m.meta, map[string]interface{}{"value": 0}, time.Now())
- if err == nil {
- y.AddTag("device", device)
- if strings.Contains(name, "byte") {
- y.AddMeta("unit", "Byte")
- }
- output <- y
- }
+ var value interface{}
+ switch def.calc {
+ case "none":
+ value = use_x
+ y, err = lp.New(def.name, m.tags, m.meta, map[string]interface{}{"value": value}, time.Now())
+ case "difference":
+ value = use_x - devData[def.name]
+ if value.(int64) < 0 {
+ value = 0
}
+ y, err = lp.New(def.name, m.tags, m.meta, map[string]interface{}{"value": value}, time.Now())
+ case "derivative":
+ value = float64(use_x-devData[def.name]) / tdiff.Seconds()
+ if value.(float64) < 0 {
+ value = 0
+ }
+ y, err = lp.New(def.name, m.tags, m.meta, map[string]interface{}{"value": value}, time.Now())
}
+ if err == nil {
+ y.AddTag("device", device)
+ if len(def.unit) > 0 {
+ y.AddMeta("unit", def.unit)
+ }
+ output <- y
+ }
+ devData[def.name] = use_x
}
}
m.lastTimestamp = now
diff --git a/collectors/lustreMetric.md b/collectors/lustreMetric.md
index de4ed60..f11b85f 100644
--- a/collectors/lustreMetric.md
+++ b/collectors/lustreMetric.md
@@ -3,32 +3,44 @@
```json
"lustrestat": {
- "procfiles" : [
- "/proc/fs/lustre/llite/lnec-XXXXXX/stats"
- ],
+ "lctl_command": "/path/to/lctl",
"exclude_metrics": [
"setattr",
"getattr"
],
"send_abs_values" : true,
- "send_derived_values" : true
+ "send_derived_values" : true,
+ "send_diff_values": true,
+ "use_sudo": false
}
```
-The `lustrestat` collector reads from the procfs stat files for Lustre like `/proc/fs/lustre/llite/lnec-XXXXXX/stats`.
+The `lustrestat` collector uses the `lctl` application with the `get_param` option to get all `llite` metrics (Lustre client). The `llite` metrics are only available for root users. If password-less sudo is configured, you can enable `sudo` in the configuration.
Metrics:
-* `lustre_read_bytes`
-* `lustre_read_requests`
-* `lustre_write_bytes`
-* `lustre_write_requests`
+* `lustre_read_bytes` (unit `bytes`)
+* `lustre_read_requests` (unit `requests`)
+* `lustre_write_bytes` (unit `bytes`)
+* `lustre_write_requests` (unit `requests`)
* `lustre_open`
* `lustre_close`
* `lustre_getattr`
* `lustre_setattr`
* `lustre_statfs`
* `lustre_inode_permission`
-* `lustre_read_bytes_bw` (if `send_derived_values == true`)
-* `lustre_write_bytes_bw` (if `send_derived_values == true`)
+* `lustre_read_bw` (if `send_derived_values == true`, unit `bytes/sec`)
+* `lustre_write_bw` (if `send_derived_values == true`, unit `bytes/sec`)
+* `lustre_read_requests_rate` (if `send_derived_values == true`, unit `requests/sec`)
+* `lustre_write_requests_rate` (if `send_derived_values == true`, unit `requests/sec`)
+* `lustre_read_bytes_diff` (if `send_diff_values == true`, unit `bytes`)
+* `lustre_read_requests_diff` (if `send_diff_values == true`, unit `requests`)
+* `lustre_write_bytes_diff` (if `send_diff_values == true`, unit `bytes`)
+* `lustre_write_requests_diff` (if `send_diff_values == true`, unit `requests`)
+* `lustre_open_diff` (if `send_diff_values == true`)
+* `lustre_close_diff` (if `send_diff_values == true`)
+* `lustre_getattr_diff` (if `send_diff_values == true`)
+* `lustre_setattr_diff` (if `send_diff_values == true`)
+* `lustre_statfs_diff` (if `send_diff_values == true`)
+* `lustre_inode_permission_diff` (if `send_diff_values == true`)
This collector adds an `device` tag.
\ No newline at end of file
diff --git a/collectors/memstatMetric.go b/collectors/memstatMetric.go
index bd7af5d..c6c7f34 100644
--- a/collectors/memstatMetric.go
+++ b/collectors/memstatMetric.go
@@ -40,8 +40,13 @@ type MemstatCollector struct {
sendMemUsed bool
}
-func getStats(filename string) map[string]float64 {
- stats := make(map[string]float64)
+type MemstatStats struct {
+ value float64
+ unit string
+}
+
+func getStats(filename string) map[string]MemstatStats {
+ stats := make(map[string]MemstatStats)
file, err := os.Open(filename)
if err != nil {
cclog.Error(err.Error())
@@ -55,12 +60,18 @@ func getStats(filename string) map[string]float64 {
if len(linefields) == 3 {
v, err := strconv.ParseFloat(linefields[1], 64)
if err == nil {
- stats[strings.Trim(linefields[0], ":")] = v
+ stats[strings.Trim(linefields[0], ":")] = MemstatStats{
+ value: v,
+ unit: linefields[2],
+ }
}
} else if len(linefields) == 5 {
v, err := strconv.ParseFloat(linefields[3], 64)
if err == nil {
- stats[strings.Trim(linefields[0], ":")] = v
+ stats[strings.Trim(linefields[0], ":")] = MemstatStats{
+ value: v,
+ unit: linefields[4],
+ }
}
}
}
@@ -78,7 +89,7 @@ func (m *MemstatCollector) Init(config json.RawMessage) error {
return err
}
}
- m.meta = map[string]string{"source": m.name, "group": "Memory", "unit": "GByte"}
+ m.meta = map[string]string{"source": m.name, "group": "Memory"}
m.stats = make(map[string]int64)
m.matches = make(map[string]string)
m.tags = map[string]string{"type": "node"}
@@ -151,30 +162,51 @@ func (m *MemstatCollector) Read(interval time.Duration, output chan lp.CCMetric)
return
}
- sendStats := func(stats map[string]float64, tags map[string]string) {
+ sendStats := func(stats map[string]MemstatStats, tags map[string]string) {
for match, name := range m.matches {
var value float64 = 0
+ var unit string = ""
if v, ok := stats[match]; ok {
- value = v
+ value = v.value
+ if len(v.unit) > 0 {
+ unit = v.unit
+ }
}
- y, err := lp.New(name, tags, m.meta, map[string]interface{}{"value": value * 1e-6}, time.Now())
+
+ y, err := lp.New(name, tags, m.meta, map[string]interface{}{"value": value}, time.Now())
if err == nil {
+ if len(unit) > 0 {
+ y.AddMeta("unit", unit)
+ }
output <- y
}
}
if m.sendMemUsed {
memUsed := 0.0
+ unit := ""
if totalVal, total := stats["MemTotal"]; total {
if freeVal, free := stats["MemFree"]; free {
if bufVal, buffers := stats["Buffers"]; buffers {
if cacheVal, cached := stats["Cached"]; cached {
- memUsed = totalVal - (freeVal + bufVal + cacheVal)
+ memUsed = totalVal.value - (freeVal.value + bufVal.value + cacheVal.value)
+ if len(totalVal.unit) > 0 {
+ unit = totalVal.unit
+ } else if len(freeVal.unit) > 0 {
+ unit = freeVal.unit
+ } else if len(bufVal.unit) > 0 {
+ unit = bufVal.unit
+ } else if len(cacheVal.unit) > 0 {
+ unit = cacheVal.unit
+ }
}
}
}
}
- y, err := lp.New("mem_used", tags, m.meta, map[string]interface{}{"value": memUsed * 1e-6}, time.Now())
+ y, err := lp.New("mem_used", tags, m.meta, map[string]interface{}{"value": memUsed}, time.Now())
if err == nil {
+ if len(unit) > 0 {
+ y.AddMeta("unit", unit)
+ }
output <- y
}
}
diff --git a/collectors/metricCollector.go b/collectors/metricCollector.go
index 747772f..7c04e90 100644
--- a/collectors/metricCollector.go
+++ b/collectors/metricCollector.go
@@ -125,5 +125,5 @@ func RemoveFromStringList(s []string, r string) ([]string, error) {
return append(s[:i], s[i+1:]...), nil
}
}
- return s, fmt.Errorf("No such string in list")
+ return s, fmt.Errorf("no such string in list")
}
diff --git a/collectors/nfsMetric.go b/collectors/nfsMetric.go
index 07e684d..c511b0d 100644
--- a/collectors/nfsMetric.go
+++ b/collectors/nfsMetric.go
@@ -36,7 +36,7 @@ type nfsCollector struct {
}
func (m *nfsCollector) initStats() error {
- cmd := exec.Command(m.config.Nfsstats, `-l`)
+ cmd := exec.Command(m.config.Nfsstats, `-l`, `--all`)
cmd.Wait()
buffer, err := cmd.Output()
if err == nil {
@@ -52,7 +52,7 @@ func (m *nfsCollector) initStats() error {
if err == nil {
x := m.data[name]
x.current = value
- x.last = 0
+ x.last = value
m.data[name] = x
}
}
@@ -63,7 +63,7 @@ func (m *nfsCollector) initStats() error {
}
func (m *nfsCollector) updateStats() error {
- cmd := exec.Command(m.config.Nfsstats, `-l`)
+ cmd := exec.Command(m.config.Nfsstats, `-l`, `--all`)
cmd.Wait()
buffer, err := cmd.Output()
if err == nil {
diff --git a/collectors/tempMetric.go b/collectors/tempMetric.go
index bbc5100..7ba8eb1 100644
--- a/collectors/tempMetric.go
+++ b/collectors/tempMetric.go
@@ -70,10 +70,10 @@ func (m *TempCollector) Init(config json.RawMessage) error {
globPattern := filepath.Join("/sys/class/hwmon", "*", "temp*_input")
inputFiles, err := filepath.Glob(globPattern)
if err != nil {
- return fmt.Errorf("Unable to glob files with pattern '%s': %v", globPattern, err)
+ return fmt.Errorf("unable to glob files with pattern '%s': %v", globPattern, err)
}
if inputFiles == nil {
- return fmt.Errorf("Unable to find any files with pattern '%s'", globPattern)
+ return fmt.Errorf("unable to find any files with pattern '%s'", globPattern)
}
// Get sensor name for each temperature sensor file
@@ -158,7 +158,7 @@ func (m *TempCollector) Init(config json.RawMessage) error {
// Empty sensors map
if len(m.sensors) == 0 {
- return fmt.Errorf("No temperature sensors found")
+ return fmt.Errorf("no temperature sensors found")
}
// Finished initialization
diff --git a/collectors/topprocsMetric.go b/collectors/topprocsMetric.go
index dd6bff3..408c3cc 100644
--- a/collectors/topprocsMetric.go
+++ b/collectors/topprocsMetric.go
@@ -39,14 +39,14 @@ func (m *TopProcsCollector) Init(config json.RawMessage) error {
m.config.Num_procs = int(DEFAULT_NUM_PROCS)
}
if m.config.Num_procs <= 0 || m.config.Num_procs > MAX_NUM_PROCS {
- return errors.New(fmt.Sprintf("num_procs option must be set in 'topprocs' config (range: 1-%d)", MAX_NUM_PROCS))
+ return fmt.Errorf("num_procs option must be set in 'topprocs' config (range: 1-%d)", MAX_NUM_PROCS)
}
m.setup()
command := exec.Command("ps", "-Ao", "comm", "--sort=-pcpu")
command.Wait()
_, err = command.Output()
if err != nil {
- return errors.New("Failed to execute command")
+ return errors.New("failed to execute command")
}
m.init = true
return nil
diff --git a/go.mod b/go.mod
index 130f5cc..07d46f6 100644
--- a/go.mod
+++ b/go.mod
@@ -3,17 +3,14 @@ module github.com/ClusterCockpit/cc-metric-collector
go 1.16
require (
- github.com/NVIDIA/go-nvml v0.11.1-0
- github.com/influxdata/influxdb-client-go/v2 v2.7.0
- github.com/influxdata/line-protocol v0.0.0-20210922203350-b1ad95c89adf
- github.com/nats-io/nats.go v1.13.1-0.20211122170419-d7c1d78a50fc
- golang.org/x/sys v0.0.0-20220114195835-da31bd327af9
- gopkg.in/Knetic/govaluate.v2 v2.3.0
-)
-
-require (
+ github.com/NVIDIA/go-nvml v0.11.6-0
github.com/PaesslerAG/gval v1.1.2
- github.com/golang/protobuf v1.5.2 // indirect
- github.com/nats-io/nats-server/v2 v2.7.0 // indirect
- google.golang.org/protobuf v1.27.1 // indirect
+ github.com/gorilla/mux v1.8.0
+ github.com/influxdata/influxdb-client-go/v2 v2.8.1
+ github.com/influxdata/line-protocol v0.0.0-20210922203350-b1ad95c89adf
+ github.com/nats-io/nats-server/v2 v2.8.0 // indirect
+ github.com/nats-io/nats.go v1.14.0
+ github.com/prometheus/client_golang v1.12.1
+ github.com/stmcginnis/gofish v0.13.0
+ golang.org/x/sys v0.0.0-20220412211240-33da011f77ad
)
diff --git a/internal/ccMetric/ccMetric.go b/internal/ccMetric/ccMetric.go
index 1de325a..661b9a4 100644
--- a/internal/ccMetric/ccMetric.go
+++ b/internal/ccMetric/ccMetric.go
@@ -24,8 +24,8 @@ type ccMetric struct {
// ccMetric access functions
type CCMetric interface {
- ToPoint(metaAsTags bool) *write.Point // Generate influxDB point for data type ccMetric
- ToLineProtocol(metaAsTags bool) string // Generate influxDB line protocol for data type ccMetric
+ ToPoint(metaAsTags map[string]bool) *write.Point // Generate influxDB point for data type ccMetric
+ ToLineProtocol(metaAsTags map[string]bool) string // Generate influxDB line protocol for data type ccMetric
Name() string // Get metric name
SetName(name string) // Set metric name
@@ -61,25 +61,18 @@ func (m *ccMetric) String() string {
}
// ToLineProtocol generates influxDB line protocol for data type ccMetric
-func (m *ccMetric) ToPoint(metaAsTags bool) (p *write.Point) {
-
- if !metaAsTags {
- p = influxdb2.NewPoint(m.name, m.tags, m.fields, m.tm)
- } else {
- tags := make(map[string]string, len(m.tags)+len(m.meta))
- for key, value := range m.tags {
- tags[key] = value
+func (m *ccMetric) ToPoint(metaAsTags map[string]bool) (p *write.Point) {
+ p = influxdb2.NewPoint(m.name, m.tags, m.fields, m.tm)
+ for key, ok1 := range metaAsTags {
+ if val, ok2 := m.GetMeta(key); ok1 && ok2 {
+ p.AddTag(key, val)
}
- for key, value := range m.meta {
- tags[key] = value
- }
- p = influxdb2.NewPoint(m.name, tags, m.fields, m.tm)
}
- return
+ return p
}
// ToLineProtocol generates influxDB line protocol for data type ccMetric
-func (m *ccMetric) ToLineProtocol(metaAsTags bool) string {
+func (m *ccMetric) ToLineProtocol(metaAsTags map[string]bool) string {
return write.PointToLineProtocol(
m.ToPoint(metaAsTags),
diff --git a/internal/ccTopology/ccTopology.go b/internal/ccTopology/ccTopology.go
index 958bb45..f68c3f4 100644
--- a/internal/ccTopology/ccTopology.go
+++ b/internal/ccTopology/ccTopology.go
@@ -169,7 +169,10 @@ func DieList() []int {
}
}
}
- return dielist
+ if len(dielist) > 0 {
+ return dielist
+ }
+ return SocketList()
}
type CpuEntry struct {
@@ -261,7 +264,7 @@ func CpuData() []CpuEntry {
for _, c := range CpuList() {
clist = append(clist, CpuEntry{Cpuid: c})
}
- for _, centry := range clist {
+ for i, centry := range clist {
centry.Socket = -1
centry.Numadomain = -1
centry.Die = -1
@@ -289,6 +292,8 @@ func CpuData() []CpuEntry {
// Lookup NUMA domain id
centry.Numadomain = getNumaDomain(base)
+ // Update values in output list
+ clist[i] = centry
}
return clist
}
diff --git a/internal/metricRouter/README.md b/internal/metricRouter/README.md
index ea5469d..adfb40b 100644
--- a/internal/metricRouter/README.md
+++ b/internal/metricRouter/README.md
@@ -8,7 +8,8 @@ The CCMetric router sits in between the collectors and the sinks and can be used
{
"num_cache_intervals" : 1,
"interval_timestamp" : true,
- "normalize_units": true,
+ "hostname_tag" : "hostname",
+ "max_forward" : 50,
"add_tags" : [
{
"key" : "cluster",
@@ -61,6 +62,20 @@ The CCMetric router sits in between the collectors and the sinks and can be used
```
There are three main options `add_tags`, `delete_tags` and `interval_timestamp`. `add_tags` and `delete_tags` are lists consisting of dicts with `key`, `value` and `if`. The `value` can be omitted in the `delete_tags` part as it only uses the `key` for removal. The `interval_timestamp` setting means that a unique timestamp is applied to all metrics traversing the router during an interval.
+
+# Processing order in the router
+
+- Add the `hostname_tag` tag (if sent by collectors or cache)
+- If `interval_timestamp == true`, change time of metrics
+- Check if metric should be dropped (`drop_metrics` and `drop_metrics_if`)
+- Add tags from `add_tags`
+- Delete tags from `del_tags`
+- Rename metric based on `rename_metrics` and store old name as `oldname` in meta information
+- Add tags from `add_tags` (if you used the new name in the `if` condition)
+- Delete tags from `del_tags` (if you used the new name in the `if` condition)
+- Send to sinks
+- Move to cache (if `num_cache_intervals > 0`)
+
# The `interval_timestamp` option
The collectors' `Read()` functions are not called simultaneously and therefore the metrics gathered in an interval can have different timestamps. If you want to avoid that and have a common timestamp (the beginning of the interval), set this option to `true` and the MetricRouter sets the time.
@@ -71,6 +86,14 @@ If the MetricRouter should buffer metrics of intervals in a MetricCache, this op
A `num_cache_intervals > 0` is required to use the `interval_aggregates` option.
+# The `hostname_tag` option
+
+By default, the router tags metrics with the hostname for all locally created metrics. The default tag name is `hostname`, but it can be changed if your organization wants anything else
+
+# The `max_forward` option
+
+Every time the router receives a metric through any of the channels, it tries to directly read up to `max_forward` metrics from the same channel. This was done as the router thread would go to sleep and wake up with every arriving metric. The default are `50` metrics at once and `max_forward` needs to greater than `1`.
+
# The `rename_metrics` option
In the ClusterCockpit world we specified a set of standard metrics. Since some collectors determine the metric names based on files, execuables and libraries, they might change from system to system (or installation to installtion, OS to OS, ...). In order to get the common names, you can rename incoming metrics before sending them to the sink. If the metric name matches the `oldname`, it is changed to `newname`
diff --git a/internal/metricRouter/metricRouter.go b/internal/metricRouter/metricRouter.go
index 330c02c..8564f24 100644
--- a/internal/metricRouter/metricRouter.go
+++ b/internal/metricRouter/metricRouter.go
@@ -51,7 +51,6 @@ type metricRouter struct {
done chan bool // channel to finish / stop metric router
wg *sync.WaitGroup // wait group for all goroutines in cc-metric-collector
timestamp time.Time // timestamp periodically updated by ticker each interval
- timerdone chan bool // channel to finish / stop timestamp updater
ticker mct.MultiChanTicker // periodically ticking once each interval
config metricRouterConfig // json encoded config for metric router
cache MetricCache // pointer to MetricCache
@@ -106,7 +105,10 @@ func (r *metricRouter) Init(ticker mct.MultiChanTicker, wg *sync.WaitGroup, rout
cclog.ComponentError("MetricRouter", err.Error())
return err
}
- r.maxForward = r.config.MaxForward
+ r.maxForward = 1
+ if r.config.MaxForward > r.maxForward {
+ r.maxForward = r.config.MaxForward
+ }
if r.config.NumCacheIntervals > 0 {
r.cache, err = NewCache(r.cache_input, r.ticker, &r.cachewg, r.config.NumCacheIntervals)
if err != nil {
@@ -124,29 +126,6 @@ func (r *metricRouter) Init(ticker mct.MultiChanTicker, wg *sync.WaitGroup, rout
return nil
}
-// StartTimer starts a timer which updates timestamp periodically
-func (r *metricRouter) StartTimer() {
- m := make(chan time.Time)
- r.ticker.AddChannel(m)
- r.timerdone = make(chan bool)
-
- r.wg.Add(1)
- go func() {
- defer r.wg.Done()
- for {
- select {
- case <-r.timerdone:
- close(r.timerdone)
- cclog.ComponentDebug("MetricRouter", "TIMER DONE")
- return
- case t := <-m:
- r.timestamp = t
- }
- }
- }()
- cclog.ComponentDebug("MetricRouter", "TIMER START")
-}
-
func getParamMap(point lp.CCMetric) map[string]interface{} {
params := make(map[string]interface{})
params["metric"] = point
@@ -267,8 +246,9 @@ func (r *metricRouter) prepareUnit(point lp.CCMetric) bool {
func (r *metricRouter) Start() {
// start timer if configured
r.timestamp = time.Now()
+ timeChan := make(chan time.Time)
if r.config.IntervalStamp {
- r.StartTimer()
+ r.ticker.AddChannel(timeChan)
}
// Router manager is done
@@ -350,6 +330,10 @@ func (r *metricRouter) Start() {
done()
return
+ case timestamp := <-timeChan:
+ r.timestamp = timestamp
+ cclog.ComponentDebug("MetricRouter", "Update timestamp", r.timestamp.UnixNano())
+
case p := <-r.coll_input:
coll_forward(p)
for i := 0; len(r.coll_input) > 0 && i < (r.maxForward-1); i++ {
@@ -395,14 +379,6 @@ func (r *metricRouter) Close() {
// wait for close of channel r.done
<-r.done
- // stop timer
- if r.config.IntervalStamp {
- cclog.ComponentDebug("MetricRouter", "TIMER CLOSE")
- r.timerdone <- true
- // wait for close of channel r.timerdone
- <-r.timerdone
- }
-
// stop metric cache
if r.config.NumCacheIntervals > 0 {
cclog.ComponentDebug("MetricRouter", "CACHE CLOSE")
diff --git a/receivers.json b/receivers.json
index a27f07d..cd78eb6 100644
--- a/receivers.json
+++ b/receivers.json
@@ -4,5 +4,22 @@
"address": "nats://my-url",
"port" : "4222",
"database": "testcluster"
+ },
+ "redfish_recv": {
+ "type": "redfish",
+ "client_config": [
+ {
+ "hostname": "my-host-1",
+ "username": "username-1",
+ "password": "password-1",
+ "endpoint": "https://my-endpoint-1"
+ },
+ {
+ "hostname": "my-host-2",
+ "username": "username-2",
+ "password": "password-2",
+ "endpoint": "https://my-endpoint-2"
+ }
+ ]
}
}
diff --git a/receivers/receiveManager.go b/receivers/receiveManager.go
index 1c13026..7a20fac 100644
--- a/receivers/receiveManager.go
+++ b/receivers/receiveManager.go
@@ -10,14 +10,13 @@ import (
)
var AvailableReceivers = map[string]func(name string, config json.RawMessage) (Receiver, error){
- "nats": NewNatsReceiver,
+ "nats": NewNatsReceiver,
+ "redfish": NewRedfishReceiver,
}
type receiveManager struct {
inputs []Receiver
output chan lp.CCMetric
- done chan bool
- wg *sync.WaitGroup
config []json.RawMessage
}
@@ -33,8 +32,6 @@ func (rm *receiveManager) Init(wg *sync.WaitGroup, receiverConfigFile string) er
// Initialize struct fields
rm.inputs = make([]Receiver, 0)
rm.output = nil
- rm.done = make(chan bool)
- rm.wg = wg
rm.config = make([]json.RawMessage, 0)
configFile, err := os.Open(receiverConfigFile)
@@ -58,7 +55,7 @@ func (rm *receiveManager) Init(wg *sync.WaitGroup, receiverConfigFile string) er
}
func (rm *receiveManager) Start() {
- rm.wg.Add(1)
+ cclog.ComponentDebug("ReceiveManager", "START")
for _, r := range rm.inputs {
cclog.ComponentDebug("ReceiveManager", "START", r.Name())
@@ -97,16 +94,19 @@ func (rm *receiveManager) AddOutput(output chan lp.CCMetric) {
}
func (rm *receiveManager) Close() {
+ cclog.ComponentDebug("ReceiveManager", "CLOSE")
+
+ // Close all receivers
for _, r := range rm.inputs {
cclog.ComponentDebug("ReceiveManager", "CLOSE", r.Name())
r.Close()
}
- rm.wg.Done()
- cclog.ComponentDebug("ReceiveManager", "CLOSE")
+
+ cclog.ComponentDebug("ReceiveManager", "DONE")
}
func New(wg *sync.WaitGroup, receiverConfigFile string) (ReceiveManager, error) {
- r := &receiveManager{}
+ r := new(receiveManager)
err := r.Init(wg, receiverConfigFile)
if err != nil {
return nil, err
diff --git a/receivers/redfishReceiver.go b/receivers/redfishReceiver.go
new file mode 100644
index 0000000..d50cbc7
--- /dev/null
+++ b/receivers/redfishReceiver.go
@@ -0,0 +1,324 @@
+package receivers
+
+import (
+ "encoding/json"
+ "fmt"
+ "strconv"
+ "sync"
+ "time"
+
+ cclog "github.com/ClusterCockpit/cc-metric-collector/internal/ccLogger"
+ lp "github.com/ClusterCockpit/cc-metric-collector/internal/ccMetric"
+
+ // See: https://pkg.go.dev/github.com/stmcginnis/gofish
+ "github.com/stmcginnis/gofish"
+)
+
+// RedfishReceiver configuration:
+type RedfishReceiver struct {
+ receiver
+ config struct {
+ Type string `json:"type"`
+ Fanout int `json:"fanout,omitempty"` // Default fanout: 64
+ Interval int `json:"interval,omitempty"` // Default interval: 30s
+
+ // Client config for each redfish service
+ ClientConfigs []struct {
+ Hostname *string `json:"hostname"`
+ Username *string `json:"username"`
+ Password *string `json:"password"`
+ Endpoint *string `json:"endpoint"`
+ Insecure *bool `json:"insecure,omitempty"`
+ ExcludeMetrics []string `json:"exclude_metrics,omitempty"`
+ gofish gofish.ClientConfig
+ } `json:"client_config"`
+ }
+
+ done chan bool // channel to finish / stop redfish receiver
+ wg sync.WaitGroup // wait group for redfish receiver
+}
+
+// Start starts the redfish receiver
+func (r *RedfishReceiver) Start() {
+ cclog.ComponentDebug(r.name, "START")
+
+ // readPowerMetric reads readfish power metric from the endpoint configured in conf
+ readPowerMetric := func(clientConfigIndex int) error {
+
+ clientConfig := &r.config.ClientConfigs[clientConfigIndex]
+
+ // Connect to redfish service
+ c, err := gofish.Connect(clientConfig.gofish)
+ if err != nil {
+ c := struct {
+ Username string
+ Endpoint string
+ BasicAuth bool
+ Insecure bool
+ }{
+ Username: clientConfig.gofish.Username,
+ Endpoint: clientConfig.gofish.Endpoint,
+ BasicAuth: clientConfig.gofish.BasicAuth,
+ Insecure: clientConfig.gofish.Insecure,
+ }
+ return fmt.Errorf("readPowerMetric: gofish.Connect(%+v) failed: %v", c, err)
+ }
+ defer c.Logout()
+
+ // Get all chassis managed by this service
+ chassis_list, err := c.Service.Chassis()
+ if err != nil {
+ return fmt.Errorf("readPowerMetric: c.Service.Chassis() failed: %v", err)
+ }
+
+ for _, chassis := range chassis_list {
+ timestamp := time.Now()
+
+ // Get power information for each chassis
+ power, err := chassis.Power()
+ if err != nil {
+ return fmt.Errorf("readPowerMetric: chassis.Power() failed: %v", err)
+ }
+ if power == nil {
+ continue
+ }
+
+ // Read min, max and average consumed watts for each power control
+ for _, pc := range power.PowerControl {
+
+ // Map of collected metrics
+ metrics := map[string]float32{
+ // PowerConsumedWatts shall represent the actual power being consumed (in
+ // Watts) by the chassis
+ "consumed_watts": pc.PowerConsumedWatts,
+ // AverageConsumedWatts shall represent the
+ // average power level that occurred averaged over the last IntervalInMin
+ // minutes.
+ "average_consumed_watts": pc.PowerMetrics.AverageConsumedWatts,
+ // MinConsumedWatts shall represent the
+ // minimum power level in watts that occurred within the last
+ // IntervalInMin minutes.
+ "min_consumed_watts": pc.PowerMetrics.MinConsumedWatts,
+ // MaxConsumedWatts shall represent the
+ // maximum power level in watts that occurred within the last
+ // IntervalInMin minutes
+ "max_consumed_watts": pc.PowerMetrics.MaxConsumedWatts,
+ }
+ intervalInMin := strconv.FormatFloat(float64(pc.PowerMetrics.IntervalInMin), 'f', -1, 32)
+
+ // Metrics to exclude
+ for _, key := range clientConfig.ExcludeMetrics {
+ delete(metrics, key)
+ }
+
+ // Set tags
+ tags := map[string]string{
+ "hostname": *clientConfig.Hostname,
+ "type": "node",
+ // ID uniquely identifies the resource
+ "id": pc.ID,
+ // MemberID shall uniquely identify the member within the collection. For
+ // services supporting Redfish v1.6 or higher, this value shall be the
+ // zero-based array index.
+ "member_id": pc.MemberID,
+ // PhysicalContext shall be a description of the affected device(s) or region
+ // within the chassis to which this power control applies.
+ "physical_context": string(pc.PhysicalContext),
+ // Name
+ "power_control_name": pc.Name,
+ }
+
+ // Delete empty tags
+ for key, value := range tags {
+ if value == "" {
+ delete(tags, key)
+ }
+ }
+
+ // Set meta data tags
+ meta := map[string]string{
+ "source": r.name,
+ "group": "Energy",
+ "interval_in_minutes": intervalInMin,
+ "unit": "watts",
+ }
+
+ // Delete empty meta data tags
+ for key, value := range meta {
+ if value == "" {
+ delete(meta, key)
+ }
+ }
+
+ for name, value := range metrics {
+
+ y, err := lp.New(name, tags, meta,
+ map[string]interface{}{
+ "value": value,
+ },
+ timestamp)
+ if err == nil {
+ r.sink <- y
+ }
+ }
+ }
+ }
+
+ return nil
+ }
+
+ // doReadPowerMetric read power metrics for all configure redfish services.
+ // To compensate latencies of the Redfish services a fanout is used.
+ doReadPowerMetric := func() {
+
+ // Compute fanout to use
+ realFanout := r.config.Fanout
+ if len(r.config.ClientConfigs) < realFanout {
+ realFanout = len(r.config.ClientConfigs)
+ }
+
+ // Create wait group and input channel for workers
+ var workerWaitGroup sync.WaitGroup
+ workerInput := make(chan int, realFanout)
+
+ // Create worker go routines
+ for i := 0; i < realFanout; i++ {
+ // Increment worker wait group counter
+ workerWaitGroup.Add(1)
+ go func() {
+ // Decrement worker wait group counter
+ defer workerWaitGroup.Done()
+
+ // Read power metrics for each client config
+ for clientConfigIndex := range workerInput {
+ err := readPowerMetric(clientConfigIndex)
+ if err != nil {
+ cclog.ComponentError(r.name, err)
+ }
+ }
+ }()
+ }
+
+ // Distribute client configs to workers
+ for i := range r.config.ClientConfigs {
+ // Check done channel status
+ select {
+ case workerInput <- i:
+ case <-r.done:
+ // process done event
+ // Stop workers, clear channel and wait for all workers to finish
+ close(workerInput)
+ for range workerInput {
+ }
+ workerWaitGroup.Wait()
+ return
+ }
+ }
+
+ // Stop workers and wait for all workers to finish
+ close(workerInput)
+ workerWaitGroup.Wait()
+ }
+
+ // Start redfish receiver
+ r.wg.Add(1)
+ go func() {
+ defer r.wg.Done()
+
+ // Create ticker
+ ticker := time.NewTicker(time.Duration(r.config.Interval) * time.Second)
+ defer ticker.Stop()
+
+ for {
+ doReadPowerMetric()
+
+ select {
+ case <-ticker.C:
+ // process ticker event -> continue
+ continue
+ case <-r.done:
+ // process done event
+ return
+ }
+ }
+ }()
+
+ cclog.ComponentDebug(r.name, "STARTED")
+}
+
+// Close redfish receiver
+func (r *RedfishReceiver) Close() {
+ cclog.ComponentDebug(r.name, "CLOSE")
+
+ // Send the signal and wait
+ close(r.done)
+ r.wg.Wait()
+
+ cclog.ComponentDebug(r.name, "DONE")
+}
+
+// New function to create a new instance of the receiver
+// Initialize the receiver by giving it a name and reading in the config JSON
+func NewRedfishReceiver(name string, config json.RawMessage) (Receiver, error) {
+ r := new(RedfishReceiver)
+
+ // Set name
+ r.name = fmt.Sprintf("RedfishReceiver(%s)", name)
+
+ // Create done channel
+ r.done = make(chan bool)
+
+ // Set defaults in r.config
+ // Allow overwriting these defaults by reading config JSON
+ r.config.Fanout = 64
+ r.config.Interval = 30
+
+ // Read the redfish receiver specific JSON config
+ if len(config) > 0 {
+ err := json.Unmarshal(config, &r.config)
+ if err != nil {
+ cclog.ComponentError(r.name, "Error reading config:", err.Error())
+ return nil, err
+ }
+ }
+
+ // Create gofish client config
+ for i := range r.config.ClientConfigs {
+ clientConfig := &r.config.ClientConfigs[i]
+ gofishConfig := &clientConfig.gofish
+
+ if clientConfig.Hostname == nil {
+ err := fmt.Errorf("client config number %v requires hostname", i)
+ cclog.ComponentError(r.name, err)
+ return nil, err
+ }
+
+ if clientConfig.Endpoint == nil {
+ err := fmt.Errorf("client config number %v requires endpoint", i)
+ cclog.ComponentError(r.name, err)
+ return nil, err
+ }
+ gofishConfig.Endpoint = *clientConfig.Endpoint
+
+ if clientConfig.Username == nil {
+ err := fmt.Errorf("client config number %v requires username", i)
+ cclog.ComponentError(r.name, err)
+ return nil, err
+ }
+ gofishConfig.Username = *clientConfig.Username
+
+ if clientConfig.Password == nil {
+ err := fmt.Errorf("client config number %v requires password", i)
+ cclog.ComponentError(r.name, err)
+ return nil, err
+ }
+ gofishConfig.Password = *clientConfig.Password
+
+ gofishConfig.Insecure = true
+ if clientConfig.Insecure != nil {
+ gofishConfig.Insecure = *clientConfig.Insecure
+ }
+ }
+
+ return r, nil
+}
diff --git a/receivers/sampleReceiver.go b/receivers/sampleReceiver.go
index 2892d56..19d6f25 100644
--- a/receivers/sampleReceiver.go
+++ b/receivers/sampleReceiver.go
@@ -36,16 +36,26 @@ func (r *SampleReceiver) Start() {
// or use own go routine but always make sure it exits
// as soon as it gets the signal of the r.done channel
+ //
+ // r.done = make(chan bool)
// r.wg.Add(1)
// go func() {
- // for {
- // select {
- // case <-r.done:
- // r.wg.Done()
- // return
- // }
- // }
- // r.wg.Done()
+ // defer r.wg.Done()
+ //
+ // // Create ticker
+ // ticker := time.NewTicker(30 * time.Second)
+ // defer ticker.Stop()
+ //
+ // for {
+ // readMetric()
+ // select {
+ // case <-ticker.C:
+ // // process ticker event -> continue
+ // continue
+ // case <-r.done:
+ // return
+ // }
+ // }
// }()
}
diff --git a/sinks.json b/sinks.json
index 2fdae5a..3e9be6d 100644
--- a/sinks.json
+++ b/sinks.json
@@ -1,6 +1,8 @@
{
- "mystdout" : {
- "type" : "stdout",
- "meta_as_tags" : true
+ "mystdout": {
+ "type": "stdout",
+ "meta_as_tags": [
+ "unit"
+ ]
}
-}
+}
\ No newline at end of file
diff --git a/sinks/httpSink.go b/sinks/httpSink.go
index c2dd2ea..7713638 100644
--- a/sinks/httpSink.go
+++ b/sinks/httpSink.go
@@ -42,24 +42,25 @@ func (s *HttpSink) Write(m lp.CCMetric) error {
if s.buffer.Len() == 0 && s.flushDelay != 0 {
// This is the first write since the last flush, start the flushTimer!
if s.flushTimer != nil && s.flushTimer.Stop() {
- cclog.ComponentDebug("HttpSink", "unexpected: the flushTimer was already running?")
+ cclog.ComponentDebug(s.name, "unexpected: the flushTimer was already running?")
}
// Run a batched flush for all lines that have arrived in the last second
s.flushTimer = time.AfterFunc(s.flushDelay, func() {
if err := s.Flush(); err != nil {
- cclog.ComponentError("HttpSink", "flush failed:", err.Error())
+ cclog.ComponentError(s.name, "flush failed:", err.Error())
}
})
}
- p := m.ToPoint(s.config.MetaAsTags)
+ p := m.ToPoint(s.meta_as_tags)
s.lock.Lock()
_, err := s.encoder.Encode(p)
s.lock.Unlock() // defer does not work here as Flush() takes the lock as well
if err != nil {
+ cclog.ComponentError(s.name, "encoding failed:", err.Error())
return err
}
@@ -84,6 +85,7 @@ func (s *HttpSink) Flush() error {
// Create new request to send buffer
req, err := http.NewRequest(http.MethodPost, s.config.URL, s.buffer)
if err != nil {
+ cclog.ComponentError(s.name, "failed to create request:", err.Error())
return err
}
@@ -100,12 +102,15 @@ func (s *HttpSink) Flush() error {
// Handle transport/tcp errors
if err != nil {
+ cclog.ComponentError(s.name, "transport/tcp error:", err.Error())
return err
}
// Handle application errors
if res.StatusCode != http.StatusOK {
- return errors.New(res.Status)
+ err = errors.New(res.Status)
+ cclog.ComponentError(s.name, "application error:", err.Error())
+ return err
}
return nil
@@ -114,7 +119,7 @@ func (s *HttpSink) Flush() error {
func (s *HttpSink) Close() {
s.flushTimer.Stop()
if err := s.Flush(); err != nil {
- cclog.ComponentError("HttpSink", "flush failed:", err.Error())
+ cclog.ComponentError(s.name, "flush failed:", err.Error())
}
s.client.CloseIdleConnections()
}
@@ -159,6 +164,11 @@ func NewHttpSink(name string, config json.RawMessage) (Sink, error) {
s.flushDelay = t
}
}
+ // Create lookup map to use meta infos as tags in the output metric
+ s.meta_as_tags = make(map[string]bool)
+ for _, k := range s.config.MetaAsTags {
+ s.meta_as_tags[k] = true
+ }
tr := &http.Transport{
MaxIdleConns: s.maxIdleConns,
IdleConnTimeout: s.idleConnTimeout,
diff --git a/sinks/influxAsyncSink.go b/sinks/influxAsyncSink.go
index a2cb64a..31d127d 100644
--- a/sinks/influxAsyncSink.go
+++ b/sinks/influxAsyncSink.go
@@ -6,12 +6,14 @@ import (
"encoding/json"
"errors"
"fmt"
+ "strings"
"time"
cclog "github.com/ClusterCockpit/cc-metric-collector/internal/ccLogger"
lp "github.com/ClusterCockpit/cc-metric-collector/internal/ccMetric"
influxdb2 "github.com/influxdata/influxdb-client-go/v2"
influxdb2Api "github.com/influxdata/influxdb-client-go/v2/api"
+ influxdb2ApiHttp "github.com/influxdata/influxdb-client-go/v2/api/http"
)
type InfluxAsyncSinkConfig struct {
@@ -28,10 +30,12 @@ type InfluxAsyncSinkConfig struct {
BatchSize uint `json:"batch_size,omitempty"`
// Interval, in ms, in which is buffer flushed if it has not been already written (by reaching batch size) . Default 1000ms
FlushInterval uint `json:"flush_interval,omitempty"`
- InfluxRetryInterval string `json:"retry_interval"`
- InfluxExponentialBase uint `json:"retry_exponential_base"`
- InfluxMaxRetries uint `json:"max_retries"`
- InfluxMaxRetryTime string `json:"max_retry_time"`
+ InfluxRetryInterval string `json:"retry_interval,omitempty"`
+ InfluxExponentialBase uint `json:"retry_exponential_base,omitempty"`
+ InfluxMaxRetries uint `json:"max_retries,omitempty"`
+ InfluxMaxRetryTime string `json:"max_retry_time,omitempty"`
+ CustomFlushInterval string `json:"custom_flush_interval,omitempty"`
+ MaxRetryAttempts uint `json:"max_retry_attempts,omitempty"`
}
type InfluxAsyncSink struct {
@@ -42,6 +46,8 @@ type InfluxAsyncSink struct {
config InfluxAsyncSinkConfig
influxRetryInterval uint
influxMaxRetryTime uint
+ customFlushInterval time.Duration
+ flushTimer *time.Timer
}
func (s *InfluxAsyncSink) connect() error {
@@ -60,20 +66,34 @@ func (s *InfluxAsyncSink) connect() error {
cclog.ComponentDebug(s.name, "Using URI", uri, "Org", s.config.Organization, "Bucket", s.config.Database)
clientOptions := influxdb2.DefaultOptions()
if s.config.BatchSize != 0 {
+ cclog.ComponentDebug(s.name, "Batch size", s.config.BatchSize)
clientOptions.SetBatchSize(s.config.BatchSize)
}
if s.config.FlushInterval != 0 {
+ cclog.ComponentDebug(s.name, "Flush interval", s.config.FlushInterval)
clientOptions.SetFlushInterval(s.config.FlushInterval)
}
+ if s.influxRetryInterval != 0 {
+ cclog.ComponentDebug(s.name, "MaxRetryInterval", s.influxRetryInterval)
+ clientOptions.SetMaxRetryInterval(s.influxRetryInterval)
+ }
+ if s.influxMaxRetryTime != 0 {
+ cclog.ComponentDebug(s.name, "MaxRetryTime", s.influxMaxRetryTime)
+ clientOptions.SetMaxRetryTime(s.influxMaxRetryTime)
+ }
+ if s.config.InfluxExponentialBase != 0 {
+ cclog.ComponentDebug(s.name, "Exponential Base", s.config.InfluxExponentialBase)
+ clientOptions.SetExponentialBase(s.config.InfluxExponentialBase)
+ }
+ if s.config.InfluxMaxRetries != 0 {
+ cclog.ComponentDebug(s.name, "Max Retries", s.config.InfluxMaxRetries)
+ clientOptions.SetMaxRetries(s.config.InfluxMaxRetries)
+ }
clientOptions.SetTLSConfig(
&tls.Config{
InsecureSkipVerify: true,
},
- )
- clientOptions.SetMaxRetryInterval(s.influxRetryInterval)
- clientOptions.SetMaxRetryTime(s.influxMaxRetryTime)
- clientOptions.SetExponentialBase(s.config.InfluxExponentialBase)
- clientOptions.SetMaxRetries(s.config.InfluxMaxRetries)
+ ).SetPrecision(time.Second)
s.client = influxdb2.NewClientWithOptions(uri, auth, clientOptions)
s.writeApi = s.client.WriteAPI(s.config.Organization, s.config.Database)
@@ -84,18 +104,35 @@ func (s *InfluxAsyncSink) connect() error {
if !ok {
return fmt.Errorf("connection to %s not healthy", uri)
}
+ s.writeApi.SetWriteFailedCallback(func(batch string, err influxdb2ApiHttp.Error, retryAttempts uint) bool {
+ mlist := strings.Split(batch, "\n")
+ cclog.ComponentError(s.name, fmt.Sprintf("Failed to write batch with %d metrics %d times (max: %d): %s", len(mlist), retryAttempts, s.config.MaxRetryAttempts, err.Error()))
+ return retryAttempts <= s.config.MaxRetryAttempts
+ })
return nil
}
func (s *InfluxAsyncSink) Write(m lp.CCMetric) error {
+ if s.customFlushInterval != 0 && s.flushTimer == nil {
+ // Run a batched flush for all lines that have arrived in the defined interval
+ s.flushTimer = time.AfterFunc(s.customFlushInterval, func() {
+ if err := s.Flush(); err != nil {
+ cclog.ComponentError(s.name, "flush failed:", err.Error())
+ }
+ })
+ }
s.writeApi.WritePoint(
- m.ToPoint(s.config.MetaAsTags),
+ m.ToPoint(s.meta_as_tags),
)
return nil
}
func (s *InfluxAsyncSink) Flush() error {
+ cclog.ComponentDebug(s.name, "Flushing")
s.writeApi.Flush()
+ if s.customFlushInterval != 0 && s.flushTimer != nil {
+ s.flushTimer = nil
+ }
return nil
}
@@ -110,13 +147,17 @@ func NewInfluxAsyncSink(name string, config json.RawMessage) (Sink, error) {
s.name = fmt.Sprintf("InfluxSink(%s)", name)
// Set default for maximum number of points sent to server in single request.
- s.config.BatchSize = 100
- s.influxRetryInterval = uint(time.Duration(1) * time.Second)
- s.config.InfluxRetryInterval = "1s"
- s.influxMaxRetryTime = uint(7 * time.Duration(24) * time.Hour)
- s.config.InfluxMaxRetryTime = "168h"
- s.config.InfluxMaxRetries = 20
- s.config.InfluxExponentialBase = 2
+ s.config.BatchSize = 0
+ s.influxRetryInterval = 0
+ //s.config.InfluxRetryInterval = "1s"
+ s.influxMaxRetryTime = 0
+ //s.config.InfluxMaxRetryTime = "168h"
+ s.config.InfluxMaxRetries = 0
+ s.config.InfluxExponentialBase = 0
+ s.config.FlushInterval = 0
+ s.config.CustomFlushInterval = ""
+ s.customFlushInterval = time.Duration(0)
+ s.config.MaxRetryAttempts = 1
// Default retry intervals (in seconds)
// 1 2
@@ -152,6 +193,11 @@ func NewInfluxAsyncSink(name string, config json.RawMessage) (Sink, error) {
len(s.config.Password) == 0 {
return nil, errors.New("not all configuration variables set required by InfluxAsyncSink")
}
+ // Create lookup map to use meta infos as tags in the output metric
+ s.meta_as_tags = make(map[string]bool)
+ for _, k := range s.config.MetaAsTags {
+ s.meta_as_tags[k] = true
+ }
toUint := func(duration string, def uint) uint {
t, err := time.ParseDuration(duration)
@@ -163,6 +209,15 @@ func NewInfluxAsyncSink(name string, config json.RawMessage) (Sink, error) {
s.influxRetryInterval = toUint(s.config.InfluxRetryInterval, s.influxRetryInterval)
s.influxMaxRetryTime = toUint(s.config.InfluxMaxRetryTime, s.influxMaxRetryTime)
+ // Use a own timer for calling Flush()
+ if len(s.config.CustomFlushInterval) > 0 {
+ t, err := time.ParseDuration(s.config.CustomFlushInterval)
+ if err != nil {
+ return nil, fmt.Errorf("invalid duration in 'custom_flush_interval': %v", err)
+ }
+ s.customFlushInterval = t
+ }
+
// Connect to InfluxDB server
if err := s.connect(); err != nil {
return nil, fmt.Errorf("unable to connect: %v", err)
diff --git a/sinks/influxSink.go b/sinks/influxSink.go
index ed3bb09..b382c38 100644
--- a/sinks/influxSink.go
+++ b/sinks/influxSink.go
@@ -6,28 +6,32 @@ import (
"encoding/json"
"errors"
"fmt"
+ "sync"
"time"
cclog "github.com/ClusterCockpit/cc-metric-collector/internal/ccLogger"
lp "github.com/ClusterCockpit/cc-metric-collector/internal/ccMetric"
influxdb2 "github.com/influxdata/influxdb-client-go/v2"
influxdb2Api "github.com/influxdata/influxdb-client-go/v2/api"
+ "github.com/influxdata/influxdb-client-go/v2/api/write"
)
type InfluxSinkConfig struct {
defaultSinkConfig
- Host string `json:"host,omitempty"`
- Port string `json:"port,omitempty"`
- Database string `json:"database,omitempty"`
- User string `json:"user,omitempty"`
- Password string `json:"password,omitempty"`
- Organization string `json:"organization,omitempty"`
- SSL bool `json:"ssl,omitempty"`
- RetentionPol string `json:"retention_policy,omitempty"`
- InfluxRetryInterval string `json:"retry_interval"`
- InfluxExponentialBase uint `json:"retry_exponential_base"`
- InfluxMaxRetries uint `json:"max_retries"`
- InfluxMaxRetryTime string `json:"max_retry_time"`
+ Host string `json:"host,omitempty"`
+ Port string `json:"port,omitempty"`
+ Database string `json:"database,omitempty"`
+ User string `json:"user,omitempty"`
+ Password string `json:"password,omitempty"`
+ Organization string `json:"organization,omitempty"`
+ SSL bool `json:"ssl,omitempty"`
+ FlushDelay string `json:"flush_delay,omitempty"`
+ BatchSize int `json:"batch_size,omitempty"`
+ RetentionPol string `json:"retention_policy,omitempty"`
+ // InfluxRetryInterval string `json:"retry_interval"`
+ // InfluxExponentialBase uint `json:"retry_exponential_base"`
+ // InfluxMaxRetries uint `json:"max_retries"`
+ // InfluxMaxRetryTime string `json:"max_retry_time"`
//InfluxMaxRetryDelay string `json:"max_retry_delay"` // It is mentioned in the docs but there is no way to set it
}
@@ -38,37 +42,71 @@ type InfluxSink struct {
config InfluxSinkConfig
influxRetryInterval uint
influxMaxRetryTime uint
+ batch []*write.Point
+ flushTimer *time.Timer
+ flushDelay time.Duration
+ lock sync.Mutex // Flush() runs in another goroutine, so this lock has to protect the buffer
//influxMaxRetryDelay uint
}
+// connect connects to the InfluxDB server
func (s *InfluxSink) connect() error {
- var auth string
+
+ // URI options:
+ // * http://host:port
+ // * https://host:port
var uri string
if s.config.SSL {
uri = fmt.Sprintf("https://%s:%s", s.config.Host, s.config.Port)
} else {
uri = fmt.Sprintf("http://%s:%s", s.config.Host, s.config.Port)
}
+
+ // Authentication options:
+ // * token
+ // * username:password
+ var auth string
if len(s.config.User) == 0 {
auth = s.config.Password
} else {
auth = fmt.Sprintf("%s:%s", s.config.User, s.config.Password)
}
cclog.ComponentDebug(s.name, "Using URI", uri, "Org", s.config.Organization, "Bucket", s.config.Database)
+
+ // Set influxDB client options
clientOptions := influxdb2.DefaultOptions()
+
+ // if s.influxRetryInterval != 0 {
+ // cclog.ComponentDebug(s.name, "MaxRetryInterval", s.influxRetryInterval)
+ // clientOptions.SetMaxRetryInterval(s.influxRetryInterval)
+ // }
+ // if s.influxMaxRetryTime != 0 {
+ // cclog.ComponentDebug(s.name, "MaxRetryTime", s.influxMaxRetryTime)
+ // clientOptions.SetMaxRetryTime(s.influxMaxRetryTime)
+ // }
+ // if s.config.InfluxExponentialBase != 0 {
+ // cclog.ComponentDebug(s.name, "Exponential Base", s.config.InfluxExponentialBase)
+ // clientOptions.SetExponentialBase(s.config.InfluxExponentialBase)
+ // }
+ // if s.config.InfluxMaxRetries != 0 {
+ // cclog.ComponentDebug(s.name, "Max Retries", s.config.InfluxMaxRetries)
+ // clientOptions.SetMaxRetries(s.config.InfluxMaxRetries)
+ // }
+
+ // Do not check InfluxDB certificate
clientOptions.SetTLSConfig(
&tls.Config{
InsecureSkipVerify: true,
},
)
- clientOptions.SetMaxRetryInterval(s.influxRetryInterval)
- clientOptions.SetMaxRetryTime(s.influxMaxRetryTime)
- clientOptions.SetExponentialBase(s.config.InfluxExponentialBase)
- clientOptions.SetMaxRetries(s.config.InfluxMaxRetries)
+ clientOptions.SetPrecision(time.Second)
+ // Create new writeAPI
s.client = influxdb2.NewClientWithOptions(uri, auth, clientOptions)
s.writeApi = s.client.WriteAPIBlocking(s.config.Organization, s.config.Database)
+
+ // Check InfluxDB server accessibility
ok, err := s.client.Ping(context.Background())
if err != nil {
return err
@@ -80,56 +118,142 @@ func (s *InfluxSink) connect() error {
}
func (s *InfluxSink) Write(m lp.CCMetric) error {
- err :=
- s.writeApi.WritePoint(
- context.Background(),
- m.ToPoint(s.config.MetaAsTags),
- )
- return err
+
+ if len(s.batch) == 0 && s.flushDelay != 0 {
+ // This is the first write since the last flush, start the flushTimer!
+ if s.flushTimer != nil && s.flushTimer.Stop() {
+ cclog.ComponentDebug(s.name, "unexpected: the flushTimer was already running?")
+ }
+
+ // Run a batched flush for all lines that have arrived in the last flush delay interval
+ s.flushTimer = time.AfterFunc(s.flushDelay, func() {
+ if err := s.Flush(); err != nil {
+ cclog.ComponentError(s.name, "flush failed:", err.Error())
+ }
+ })
+ }
+
+ // Append metric to batch slice
+ p := m.ToPoint(s.meta_as_tags)
+ s.lock.Lock()
+ s.batch = append(s.batch, p)
+ s.lock.Unlock()
+
+ // Flush synchronously if "flush_delay" is zero
+ if s.flushDelay == 0 {
+ return s.Flush()
+ }
+
+ // Flush if batch size is reached
+ if len(s.batch) == s.config.BatchSize {
+ return s.Flush()
+ }
+
+ return nil
}
+// Flush sends all metrics buffered in batch slice to InfluxDB server
func (s *InfluxSink) Flush() error {
+
+ // Lock access to batch slice
+ s.lock.Lock()
+ defer s.lock.Unlock()
+
+ // Nothing to do, batch slice is empty
+ if len(s.batch) == 0 {
+ return nil
+ }
+
+ // Send metrics from batch slice
+ err := s.writeApi.WritePoint(context.Background(), s.batch...)
+ if err != nil {
+ cclog.ComponentError(s.name, "flush failed:", err.Error())
+ return err
+ }
+
+ // Clear batch slice
+ for i := range s.batch {
+ s.batch[i] = nil
+ }
+ s.batch = s.batch[:0]
+
return nil
}
func (s *InfluxSink) Close() {
cclog.ComponentDebug(s.name, "Closing InfluxDB connection")
+ s.flushTimer.Stop()
+ s.Flush()
s.client.Close()
}
+// NewInfluxSink create a new InfluxDB sink
func NewInfluxSink(name string, config json.RawMessage) (Sink, error) {
s := new(InfluxSink)
s.name = fmt.Sprintf("InfluxSink(%s)", name)
+
+ // Set config default values
+ s.config.BatchSize = 100
+ s.config.FlushDelay = "1s"
+
+ // Read config
if len(config) > 0 {
err := json.Unmarshal(config, &s.config)
if err != nil {
return nil, err
}
}
- s.influxRetryInterval = uint(time.Duration(1) * time.Second)
- s.config.InfluxRetryInterval = "1s"
- s.influxMaxRetryTime = uint(7 * time.Duration(24) * time.Hour)
- s.config.InfluxMaxRetryTime = "168h"
- s.config.InfluxMaxRetries = 20
- s.config.InfluxExponentialBase = 2
+ s.influxRetryInterval = 0
+ s.influxMaxRetryTime = 0
+ // s.config.InfluxRetryInterval = ""
+ // s.config.InfluxMaxRetryTime = ""
+ // s.config.InfluxMaxRetries = 0
+ // s.config.InfluxExponentialBase = 0
- if len(s.config.Host) == 0 ||
- len(s.config.Port) == 0 ||
- len(s.config.Database) == 0 ||
- len(s.config.Organization) == 0 ||
- len(s.config.Password) == 0 {
- return nil, errors.New("not all configuration variables set required by InfluxSink")
+ if len(s.config.Host) == 0 {
+ return nil, errors.New("Missing host configuration required by InfluxSink")
+ }
+ if len(s.config.Port) == 0 {
+ return nil, errors.New("Missing port configuration required by InfluxSink")
+ }
+ if len(s.config.Database) == 0 {
+ return nil, errors.New("Missing database configuration required by InfluxSink")
+ }
+ if len(s.config.Organization) == 0 {
+ return nil, errors.New("Missing organization configuration required by InfluxSink")
+ }
+ if len(s.config.Password) == 0 {
+ return nil, errors.New("Missing password configuration required by InfluxSink")
}
- toUint := func(duration string, def uint) uint {
- t, err := time.ParseDuration(duration)
+ // Create lookup map to use meta infos as tags in the output metric
+ s.meta_as_tags = make(map[string]bool)
+ for _, k := range s.config.MetaAsTags {
+ s.meta_as_tags[k] = true
+ }
+
+ // toUint := func(duration string, def uint) uint {
+ // if len(duration) > 0 {
+ // t, err := time.ParseDuration(duration)
+ // if err == nil {
+ // return uint(t.Milliseconds())
+ // }
+ // }
+ // return def
+ // }
+ // s.influxRetryInterval = toUint(s.config.InfluxRetryInterval, s.influxRetryInterval)
+ // s.influxMaxRetryTime = toUint(s.config.InfluxMaxRetryTime, s.influxMaxRetryTime)
+
+ // Configure flush delay duration
+ if len(s.config.FlushDelay) > 0 {
+ t, err := time.ParseDuration(s.config.FlushDelay)
if err == nil {
- return uint(t.Milliseconds())
+ s.flushDelay = t
}
- return def
}
- s.influxRetryInterval = toUint(s.config.InfluxRetryInterval, s.influxRetryInterval)
- s.influxMaxRetryTime = toUint(s.config.InfluxMaxRetryTime, s.influxMaxRetryTime)
+
+ // allocate batch slice
+ s.batch = make([]*write.Point, 0, s.config.BatchSize)
// Connect to InfluxDB server
if err := s.connect(); err != nil {
diff --git a/sinks/influxSink.md b/sinks/influxSink.md
index a099895..8f9ce83 100644
--- a/sinks/influxSink.md
+++ b/sinks/influxSink.md
@@ -17,10 +17,8 @@ The `influxdb` sink uses the official [InfluxDB golang client](https://pkg.go.de
"password" : "examplepw",
"organization": "myorg",
"ssl": true,
- "retry_interval" : "1s",
- "retry_exponential_base" : 2,
- "max_retries": 20,
- "max_retry_time" : "168h"
+ "flush_delay" : "1s",
+ "batch_size" : 100
}
}
```
@@ -34,9 +32,6 @@ The `influxdb` sink uses the official [InfluxDB golang client](https://pkg.go.de
- `password`: Password for basic authentification
- `organization`: Organization in the InfluxDB
- `ssl`: Use SSL connection
-- `retry_interval`: Base retry interval for failed write requests, default 1s
-- `retry_exponential_base`: The retry interval is exponentially increased with this base, default 2
-- `max_retries`: Maximal number of retry attempts
-- `max_retry_time`: Maximal time to retry failed writes, default 168h (one week)
+- `flush_delay`: Group metrics coming in to a single batch
+- `batch_size`: Maximal batch size
-For information about the calculation of the retry interval settings, see [offical influxdb-client-go documentation](https://github.com/influxdata/influxdb-client-go#handling-of-failed-async-writes)
\ No newline at end of file
diff --git a/sinks/metricSink.go b/sinks/metricSink.go
index d5356d0..c6c6860 100644
--- a/sinks/metricSink.go
+++ b/sinks/metricSink.go
@@ -5,13 +5,13 @@ import (
)
type defaultSinkConfig struct {
- MetaAsTags bool `json:"meta_as_tags,omitempty"`
- Type string `json:"type"`
+ MetaAsTags []string `json:"meta_as_tags,omitempty"`
+ Type string `json:"type"`
}
type sink struct {
- meta_as_tags bool // Use meta data tags as tags
- name string // Name of the sink
+ meta_as_tags map[string]bool // Use meta data tags as tags
+ name string // Name of the sink
}
type Sink interface {
diff --git a/sinks/natsSink.go b/sinks/natsSink.go
index 0d7987e..0597e9b 100644
--- a/sinks/natsSink.go
+++ b/sinks/natsSink.go
@@ -55,7 +55,7 @@ func (s *NatsSink) connect() error {
func (s *NatsSink) Write(m lp.CCMetric) error {
if s.client != nil {
- _, err := s.encoder.Encode(m.ToPoint(s.config.MetaAsTags))
+ _, err := s.encoder.Encode(m.ToPoint(s.meta_as_tags))
if err != nil {
cclog.ComponentError(s.name, "Write:", err.Error())
return err
@@ -97,6 +97,11 @@ func NewNatsSink(name string, config json.RawMessage) (Sink, error) {
len(s.config.Database) == 0 {
return nil, errors.New("not all configuration variables set required by NatsSink")
}
+ // Create lookup map to use meta infos as tags in the output metric
+ s.meta_as_tags = make(map[string]bool)
+ for _, k := range s.config.MetaAsTags {
+ s.meta_as_tags[k] = true
+ }
// Setup Influx line protocol
s.buffer = &bytes.Buffer{}
s.buffer.Grow(1025)
@@ -105,7 +110,7 @@ func NewNatsSink(name string, config json.RawMessage) (Sink, error) {
s.encoder.SetMaxLineBytes(1024)
// Setup infos for connection
if err := s.connect(); err != nil {
- return nil, fmt.Errorf("Unable to connect: %v", err)
+ return nil, fmt.Errorf("unable to connect: %v", err)
}
return s, nil
}
diff --git a/sinks/sampleSink.go b/sinks/sampleSink.go
index 3913a29..2a823e6 100644
--- a/sinks/sampleSink.go
+++ b/sinks/sampleSink.go
@@ -10,14 +10,14 @@ import (
)
type SampleSinkConfig struct {
- // defines JSON tags for 'type' and 'meta_as_tags'
+ // defines JSON tags for 'type' and 'meta_as_tags' (string list)
// See: metricSink.go
defaultSinkConfig
// Additional config options, for SampleSink
}
type SampleSink struct {
- // declares elements 'name' and 'meta_as_tags'
+ // declares elements 'name' and 'meta_as_tags' (string to bool map!)
sink
config SampleSinkConfig // entry point to the SampleSinkConfig
}
@@ -28,6 +28,7 @@ type SampleSink struct {
// Code to submit a single CCMetric to the sink
func (s *SampleSink) Write(point lp.CCMetric) error {
+ // based on s.meta_as_tags use meta infos as tags
log.Print(point)
return nil
}
@@ -62,6 +63,12 @@ func NewSampleSink(name string, config json.RawMessage) (Sink, error) {
}
}
+ // Create lookup map to use meta infos as tags in the output metric
+ s.meta_as_tags = make(map[string]bool)
+ for _, k := range s.config.MetaAsTags {
+ s.meta_as_tags[k] = true
+ }
+
// Check if all required fields in the config are set
// E.g. use 'len(s.config.Option) > 0' for string settings
diff --git a/sinks/stdoutSink.go b/sinks/stdoutSink.go
index acf2621..e091af3 100644
--- a/sinks/stdoutSink.go
+++ b/sinks/stdoutSink.go
@@ -63,7 +63,11 @@ func NewStdoutSink(name string, config json.RawMessage) (Sink, error) {
s.output = f
}
}
- s.meta_as_tags = s.config.MetaAsTags
+ // Create lookup map to use meta infos as tags in the output metric
+ s.meta_as_tags = make(map[string]bool)
+ for _, k := range s.config.MetaAsTags {
+ s.meta_as_tags[k] = true
+ }
return s, nil
}