collector/cpu: Support CPU online status

Blocked by: prometheus/procfs#644.

Signed-off-by: Pranshu Srivastava <[email protected]>

Author: rexagod

collector/cpu_linux.go

@@ -17,6 +17,7 @@
 package collector
 
 import (
+	"errors"
 	"fmt"
 	"os"
 	"path/filepath"
@@ -29,13 +30,14 @@ import (
 	"github.com/go-kit/log"
 	"github.com/go-kit/log/level"
 	"github.com/prometheus/client_golang/prometheus"
-	"github.com/prometheus/procfs"
-	"github.com/prometheus/procfs/sysfs"
+	"github.com/rexagod/procfs"
+	"github.com/rexagod/procfs/sysfs"
 	"golang.org/x/exp/maps"
 )
 
 type cpuCollector struct {
-	fs                 procfs.FS
+	procfs             procfs.FS
+	sysfs              sysfs.FS
 	cpu                *prometheus.Desc
 	cpuInfo            *prometheus.Desc
 	cpuFrequencyHz     *prometheus.Desc
@@ -45,6 +47,7 @@ type cpuCollector struct {
 	cpuCoreThrottle    *prometheus.Desc
 	cpuPackageThrottle *prometheus.Desc
 	cpuIsolated        *prometheus.Desc
+	cpuOnline          *prometheus.Desc
 	logger             log.Logger
 	cpuStats           map[int64]procfs.CPUStat
 	cpuStatsMutex      sync.Mutex
@@ -71,17 +74,17 @@ func init() {
 
 // NewCPUCollector returns a new Collector exposing kernel/system statistics.
 func NewCPUCollector(logger log.Logger) (Collector, error) {
-	fs, err := procfs.NewFS(*procPath)
+	pfs, err := procfs.NewFS(*procPath)
 	if err != nil {
 		return nil, fmt.Errorf("failed to open procfs: %w", err)
 	}
 
-	sysfs, err := sysfs.NewFS(*sysPath)
+	sfs, err := sysfs.NewFS(*sysPath)
 	if err != nil {
 		return nil, fmt.Errorf("failed to open sysfs: %w", err)
 	}
 
-	isolcpus, err := sysfs.IsolatedCPUs()
+	isolcpus, err := sfs.IsolatedCPUs()
 	if err != nil {
 		if !os.IsNotExist(err) {
 			return nil, fmt.Errorf("Unable to get isolated cpus: %w", err)
@@ -90,8 +93,9 @@ func NewCPUCollector(logger log.Logger) (Collector, error) {
 	}
 
 	c := &cpuCollector{
-		fs:  fs,
-		cpu: nodeCPUSecondsDesc,
+		procfs: pfs,
+		sysfs:  sfs,
+		cpu:    nodeCPUSecondsDesc,
 		cpuInfo: prometheus.NewDesc(
 			prometheus.BuildFQName(namespace, cpuCollectorSubsystem, "info"),
 			"CPU information from /proc/cpuinfo.",
@@ -132,6 +136,11 @@ func NewCPUCollector(logger log.Logger) (Collector, error) {
 			"Whether each core is isolated, information from /sys/devices/system/cpu/isolated.",
 			[]string{"cpu"}, nil,
 		),
+		cpuOnline: prometheus.NewDesc(
+			prometheus.BuildFQName(namespace, cpuCollectorSubsystem, "online"),
+			"CPUs that are online and being scheduled.",
+			[]string{"cpu"}, nil,
+		),
 		logger:       logger,
 		isolatedCpus: isolcpus,
 		cpuStats:     make(map[int64]procfs.CPUStat),
@@ -178,12 +187,21 @@ func (c *cpuCollector) Update(ch chan<- prometheus.Metric) error {
 	if c.isolatedCpus != nil {
 		c.updateIsolated(ch)
 	}
-	return c.updateThermalThrottle(ch)
+	err := c.updateThermalThrottle(ch)
+	if err != nil {
+		return err
+	}
+	err = c.updateOnline(ch)
+	if err != nil {
+		return err
+	}
+
+	return nil
 }
 
 // updateInfo reads /proc/cpuinfo
 func (c *cpuCollector) updateInfo(ch chan<- prometheus.Metric) error {
-	info, err := c.fs.CPUInfo()
+	info, err := c.procfs.CPUInfo()
 	if err != nil {
 		return err
 	}
@@ -334,9 +352,31 @@ func (c *cpuCollector) updateIsolated(ch chan<- prometheus.Metric) {
 	}
 }
 
+// updateOnline reads /sys/devices/system/cpu/cpu*/online through sysfs and exports online status metrics.
+func (c *cpuCollector) updateOnline(ch chan<- prometheus.Metric) error {
+	cpus, err := c.sysfs.CPUs()
+	if err != nil {
+		return err
+	}
+	// No-op if the system does not support CPU online stats.
+	cpu0 := cpus[0]
+	if _, err := cpu0.Online(); err != nil && errors.Is(err, os.ErrNotExist) {
+		return nil
+	}
+	for _, cpu := range cpus {
+		setOnline := float64(0)
+		if online, _ := cpu.Online(); online {
+			setOnline = 1
+		}
+		ch <- prometheus.MustNewConstMetric(c.cpuOnline, prometheus.GaugeValue, setOnline, cpu.Number())
+	}
+
+	return nil
+}
+
 // updateStat reads /proc/stat through procfs and exports CPU-related metrics.
 func (c *cpuCollector) updateStat(ch chan<- prometheus.Metric) error {
-	stats, err := c.fs.Stat()
+	stats, err := c.procfs.Stat()
 	if err != nil {
 		return err
 	}

collector/cpu_linux_test.go

@@ -21,7 +21,7 @@ import (
 	"testing"
 
 	"github.com/go-kit/log"
-	"github.com/prometheus/procfs"
+	"github.com/rexagod/procfs"
 )
 
 func copyStats(d, s map[int64]procfs.CPUStat) {

go.mod

@@ -27,9 +27,10 @@ require (
 	github.com/prometheus/common v0.53.0
 	github.com/prometheus/exporter-toolkit v0.11.0
 	github.com/prometheus/procfs v0.14.0
+	github.com/rexagod/procfs v0.0.0-00010101000000-000000000000
 	github.com/safchain/ethtool v0.3.0
 	golang.org/x/exp v0.0.0-20240416160154-fe59bbe5cc7f
-	golang.org/x/sys v0.19.0
+	golang.org/x/sys v0.20.0
 	howett.net/plist v1.0.1
 )
 
@@ -59,3 +60,5 @@ require (
 	google.golang.org/protobuf v1.33.0 // indirect
 	gopkg.in/yaml.v2 v2.4.0 // indirect
 )
+
+replace github.com/rexagod/procfs => ../procfs // https://github.com/prometheus/procfs/pull/644

go.sum

@@ -116,8 +116,8 @@ golang.org/x/sync v0.7.0/go.mod h1:Czt+wKu1gCyEFDUtn0jG5QVvpJ6rzVqr5aXyt9drQfk=
 golang.org/x/sys v0.0.0-20190215142949-d0b11bdaac8a/go.mod h1:STP8DvDyc/dI5b8T5hshtkjS+E42TnysNCUPdjciGhY=
 golang.org/x/sys v0.0.0-20211031064116-611d5d643895/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
 golang.org/x/sys v0.6.0/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
-golang.org/x/sys v0.19.0 h1:q5f1RH2jigJ1MoAWp2KTp3gm5zAGFUTarQZ5U386+4o=
-golang.org/x/sys v0.19.0/go.mod h1:/VUhepiaJMQUp4+oa/7Zr1D23ma6VTLIYjOOTFZPUcA=
+golang.org/x/sys v0.20.0 h1:Od9JTbYCk261bKm4M/mw7AklTlFYIa0bIp9BgSm1S8Y=
+golang.org/x/sys v0.20.0/go.mod h1:/VUhepiaJMQUp4+oa/7Zr1D23ma6VTLIYjOOTFZPUcA=
 golang.org/x/text v0.3.0/go.mod h1:NqM8EUOU14njkJ3fqMW+pc6Ldnwhi/IjpwHt7yyuwOQ=
 golang.org/x/text v0.3.2/go.mod h1:bEr9sfX3Q8Zfm5fL9x+3itogRgK3+ptLWKqgva+5dAk=
 golang.org/x/text v0.14.0 h1:ScX5w1eTa3QqT8oi6+ziP7dTV1S2+ALU0bI+0zXKWiQ=