telegraf/plugins/inputs/procstat/procstat.go

package procstat

import (
	"fmt"
	"strconv"
	"time"

	"github.com/influxdata/telegraf"
	"github.com/influxdata/telegraf/plugins/inputs"
	"github.com/shirou/gopsutil/process"
)

var (
	defaultPIDFinder = NewPgrep
	defaultProcess   = NewProc
)

type PID int32

type Procstat struct {
	PidFile     string `toml:"pid_file"`
	Exe         string
	Pattern     string
	Prefix      string
	ProcessName string
	User        string
	PidTag      bool

	pidFinder       PIDFinder
	createPIDFinder func() (PIDFinder, error)
	procs           map[PID]Process
	createProcess   func(PID) (Process, error)
}

var sampleConfig = `
  ## Must specify one of: pid_file, exe, or pattern
  ## PID file to monitor process
  pid_file = "/var/run/nginx.pid"
  ## executable name (ie, pgrep <exe>)
  # exe = "nginx"
  ## pattern as argument for pgrep (ie, pgrep -f <pattern>)
  # pattern = "nginx"
  ## user as argument for pgrep (ie, pgrep -u <user>)
  # user = "nginx"

  ## override for process_name
  ## This is optional; default is sourced from /proc/<pid>/status
  # process_name = "bar"
  ## Field name prefix
  prefix = ""
  ## comment this out if you want raw cpu_time stats
  fielddrop = ["cpu_time_*"]
  ## This is optional; moves pid into a tag instead of a field
  pid_tag = false
`

func (_ *Procstat) SampleConfig() string {
	return sampleConfig
}

func (_ *Procstat) Description() string {
	return "Monitor process cpu and memory usage"
}

func (p *Procstat) Gather(acc telegraf.Accumulator) error {
	if p.createPIDFinder == nil {
		p.createPIDFinder = defaultPIDFinder
	}
	if p.createProcess == nil {
		p.createProcess = defaultProcess
	}

	procs, err := p.updateProcesses(p.procs)
	if err != nil {
		acc.AddError(fmt.Errorf("E! Error: procstat getting process, exe: [%s] pidfile: [%s] pattern: [%s] user: [%s] %s",
			p.Exe, p.PidFile, p.Pattern, p.User, err.Error()))
	}
	p.procs = procs

	for _, proc := range p.procs {
		p.addMetrics(proc, acc)
	}

	return nil
}

// Add metrics a single Process
func (p *Procstat) addMetrics(proc Process, acc telegraf.Accumulator) {
	var prefix string
	if p.Prefix != "" {
		prefix = p.Prefix + "_"
	}

	fields := map[string]interface{}{}

	//If process_name tag is not already set, set to actual name
	if _, nameInTags := proc.Tags()["process_name"]; !nameInTags {
		name, err := proc.Name()
		if err == nil {
			proc.Tags()["process_name"] = name
		}
	}

	//If pid is not present as a tag, include it as a field.
	if _, pidInTags := proc.Tags()["pid"]; !pidInTags {
		fields["pid"] = int32(proc.PID())
	}

	numThreads, err := proc.NumThreads()
	if err == nil {
		fields[prefix+"num_threads"] = numThreads
	}

	fds, err := proc.NumFDs()
	if err == nil {
		fields[prefix+"num_fds"] = fds
	}

	ctx, err := proc.NumCtxSwitches()
	if err == nil {
		fields[prefix+"voluntary_context_switches"] = ctx.Voluntary
		fields[prefix+"involuntary_context_switches"] = ctx.Involuntary
	}

	io, err := proc.IOCounters()
	if err == nil {
		fields[prefix+"read_count"] = io.ReadCount
		fields[prefix+"write_count"] = io.WriteCount
		fields[prefix+"read_bytes"] = io.ReadBytes
		fields[prefix+"write_bytes"] = io.WriteBytes
	}

	cpu_time, err := proc.Times()
	if err == nil {
		fields[prefix+"cpu_time_user"] = cpu_time.User
		fields[prefix+"cpu_time_system"] = cpu_time.System
		fields[prefix+"cpu_time_idle"] = cpu_time.Idle
		fields[prefix+"cpu_time_nice"] = cpu_time.Nice
		fields[prefix+"cpu_time_iowait"] = cpu_time.Iowait
		fields[prefix+"cpu_time_irq"] = cpu_time.Irq
		fields[prefix+"cpu_time_soft_irq"] = cpu_time.Softirq
		fields[prefix+"cpu_time_steal"] = cpu_time.Steal
		fields[prefix+"cpu_time_stolen"] = cpu_time.Stolen
		fields[prefix+"cpu_time_guest"] = cpu_time.Guest
		fields[prefix+"cpu_time_guest_nice"] = cpu_time.GuestNice
	}

	cpu_perc, err := proc.Percent(time.Duration(0))
	if err == nil {
		fields[prefix+"cpu_usage"] = cpu_perc
	}

	mem, err := proc.MemoryInfo()
	if err == nil {
		fields[prefix+"memory_rss"] = mem.RSS
		fields[prefix+"memory_vms"] = mem.VMS
		fields[prefix+"memory_swap"] = mem.Swap
		fields[prefix+"memory_data"] = mem.Data
		fields[prefix+"memory_stack"] = mem.Stack
		fields[prefix+"memory_locked"] = mem.Locked
	}

	rlims, err := proc.RlimitUsage(true)
	if err == nil {
		for _, rlim := range rlims {
			var name string
			switch rlim.Resource {
			case process.RLIMIT_CPU:
				name = "cpu_time"
			case process.RLIMIT_DATA:
				name = "memory_data"
			case process.RLIMIT_STACK:
				name = "memory_stack"
			case process.RLIMIT_RSS:
				name = "memory_rss"
			case process.RLIMIT_NOFILE:
				name = "num_fds"
			case process.RLIMIT_MEMLOCK:
				name = "memory_locked"
			case process.RLIMIT_AS:
				name = "memory_vms"
			case process.RLIMIT_LOCKS:
				name = "file_locks"
			case process.RLIMIT_SIGPENDING:
				name = "signals_pending"
			case process.RLIMIT_NICE:
				name = "nice_priority"
			case process.RLIMIT_RTPRIO:
				name = "realtime_priority"
			default:
				continue
			}

			fields[prefix+"rlimit_"+name+"_soft"] = rlim.Soft
			fields[prefix+"rlimit_"+name+"_hard"] = rlim.Hard
			if name != "file_locks" { // gopsutil doesn't currently track the used file locks count
				fields[prefix+name] = rlim.Used
			}
		}
	}

	acc.AddFields("procstat", fields, proc.Tags())
}

// Update monitored Processes
func (p *Procstat) updateProcesses(prevInfo map[PID]Process) (map[PID]Process, error) {
	pids, tags, err := p.findPids()
	if err != nil {
		return nil, err
	}

	procs := make(map[PID]Process, len(prevInfo))

	for _, pid := range pids {
		info, ok := prevInfo[pid]
		if ok {
			procs[pid] = info
		} else {
			proc, err := p.createProcess(pid)
			if err != nil {
				// No problem; process may have ended after we found it
				continue
			}
			procs[pid] = proc

			// Add initial tags
			for k, v := range tags {
				proc.Tags()[k] = v
			}

			// Add pid tag if needed
			if p.PidTag {
				proc.Tags()["pid"] = strconv.Itoa(int(pid))
			}
			if p.ProcessName != "" {
				proc.Tags()["process_name"] = p.ProcessName
			}
		}
	}
	return procs, nil
}

// Create and return PIDGatherer lazily
func (p *Procstat) getPIDFinder() (PIDFinder, error) {
	if p.pidFinder == nil {
		f, err := p.createPIDFinder()
		if err != nil {
			return nil, err
		}
		p.pidFinder = f
	}
	return p.pidFinder, nil
}

// Get matching PIDs and their initial tags
func (p *Procstat) findPids() ([]PID, map[string]string, error) {
	var pids []PID
	var tags map[string]string
	var err error

	f, err := p.getPIDFinder()
	if err != nil {
		return nil, nil, err
	}

	if p.PidFile != "" {
		pids, err = f.PidFile(p.PidFile)
		tags = map[string]string{"pidfile": p.PidFile}
	} else if p.Exe != "" {
		pids, err = f.Pattern(p.Exe)
		tags = map[string]string{"exe": p.Exe}
	} else if p.Pattern != "" {
		pids, err = f.FullPattern(p.Pattern)
		tags = map[string]string{"pattern": p.Pattern}
	} else if p.User != "" {
		pids, err = f.Uid(p.User)
		tags = map[string]string{"user": p.User}
	} else {
		err = fmt.Errorf("Either exe, pid_file, user, or pattern has to be specified")
	}

	return pids, tags, err
}

func init() {
	inputs.Add("procstat", func() telegraf.Input {
		return &Procstat{}
	})
}