statsd_exporter VS process-exporter

It's part of the statsd_exporter.

package metrics import ( "fmt" "strings" "time" "github.com/alexcesaro/statsd" ) // Hybrid metrics type Hybrid struct { Statsd StatsdConfig Client *statsd.Client AttachLabels bool } // StatsdConfig ... type StatsdConfig struct { Host string Prefix string } // Init ... func (m *Hybrid) Init() error { // Set up the statsd client client, err := statsd.New( statsd.Address(m.Statsd.Host), statsd.Prefix(m.Statsd.Prefix), statsd.FlushPeriod(time.Second*1), ) if err != nil { return fmt.Errorf("Hybrid: Could not connect to %s: %s", m.Statsd.Host, err.Error()) } m.Client = client return nil } // Format key with labels func (m *Hybrid) formatKeyWithLabels(key string, labels map[string]string) string { // Do we want to attach labels? if m.AttachLabels == false { return key } // Use librato-style tags // metrim.name#tagName=val,tag2Name=val2:0|c // https://github.com/prometheus/statsd_exporter#tagging-extensions if len(key) == 0 { return key } formattedLabels := []string{} for labelKey, labelVal := range labels { formattedLabels = append(formattedLabels, fmt.Sprintf("%s=%s", labelKey, labelVal)) } joinedFormattedLabels := strings.TrimSpace(strings.Join(formattedLabels, ",")) if joinedFormattedLabels != "" { key = fmt.Sprintf("%s#%s", key, joinedFormattedLabels) } return strings.Trim(key, "#") } // Increment ... func (m *Hybrid) Increment(key string, labels map[string]string) { m.IncrementBy(key, 1, labels) } // IncrementBy ... func (m *Hybrid) IncrementBy(key string, by int, labels map[string]string) { updatedKey := m.formatKeyWithLabels(key, labels) if m.Client != nil { m.Client.Count(updatedKey, by) } } // Gauge ... func (m *Hybrid) Gauge(key string, value int, labels map[string]string) { updatedKey := m.formatKeyWithLabels(key, labels) if m.Client != nil { m.Client.Gauge(updatedKey, value) } }

OP is specifically asking about MacOS, but for *nix use-cases, there's the process_exporter that works to expose process level statistics. https://github.com/ncabatoff/process-exporter

However, the process_exporter is useful for some situations where software has complex subprocess management and you want to track based on the cmdline running. For example all of the various PostgreSQL worker processes.

yes. The config options are different though. Just see the README https://github.com/ncabatoff/process-exporter