sanitize label values for invalid utf-8

go.mod

@@ -11,4 +11,5 @@ require (
 	github.com/prometheus/procfs v0.0.0-20190519111021-9935e8e0588d // indirect
 	github.com/yuin/gopher-lua v0.0.0-20181214045814-db9ae37725ec // indirect
 	golang.org/x/sync v0.0.0-20190227155943-e225da77a7e6 // indirect
+	golang.org/x/text v0.3.0
 )

metric.go

@@ -5,6 +5,7 @@ import (
 	"log"
 	"strconv"
 	"strings"
+	"unicode/utf8"
 
 	"github.com/prometheus/client_golang/prometheus"
 )
@@ -45,6 +46,10 @@ func infoCollect(
 ) []prometheus.Metric {
 	var res []prometheus.Metric
 	stats := parseInfo(info)
+	validLabelValues := make([]string, len(labelValues))
+	for pos, lv := range labelValues {
+		validLabelValues[pos] = sanitizeLabelValue(lv)
+	}
 	for key, m := range metrics {
 		v, ok := stats[key]
 		if !ok {
@@ -58,12 +63,26 @@ func infoCollect(
 		}
 		res = append(
 			res,
-			prometheus.MustNewConstMetric(m.desc, m.typ, f, labelValues...),
+			prometheus.MustNewConstMetric(m.desc, m.typ, f, validLabelValues...),
 		)
 	}
 	return res
 }
 
+func sanitizeLabelValue(lv string) string {
+	if utf8.ValidString(lv) {
+		return lv
+	}
+	fixUtf := func(r rune) rune {
+		if r == utf8.RuneError {
+			return 65533
+		}
+		return r
+	}
+
+	return strings.Map(fixUtf, lv)
+}
+
 func parseInfo(s string) map[string]string {
 	r := map[string]string{}
 	for _, l := range strings.Split(s, ";") {

metric_test.go

@@ -44,6 +44,7 @@ func TestInfoCollect(t *testing.T) {
 		payload string
 		field   string
 		metric  cmetric
+		labels  []string
 		want    string
 	}
 	for n, c := range []cas{
@@ -103,9 +104,41 @@ func TestInfoCollect(t *testing.T) {
 			},
 			want: `gauge:<value:1 > `,
 		},
+		{
+			payload: "counter-1=3.14:gauge-1=6.12:flag=true:counter-2=6.66",
+			field:   "flag",
+			metric: cmetric{
+				typ: prometheus.GaugeValue,
+				desc: prometheus.NewDesc(
+					"c1",
+					"My second flag",
+					[]string{"namespace", "set"},
+					nil,
+				),
+			},
+			//labels: []string{"ns", ""},
+			labels: []string{"ns", "\xC0"},
+			want: `label:<name:"namespace" value:"ns" > label:<name:"set" value:"\357\277\275" > gauge:<value:1 > `,
+		},
+		{
+			payload: "counter-1=3.14:gauge-1=6.12:flag=true:counter-2=6.66",
+			field:   "flag",
+			metric: cmetric{
+				typ: prometheus.GaugeValue,
+				desc: prometheus.NewDesc(
+					"c1",
+					"My second flag",
+					[]string{"namespace", "set"},
+					nil,
+				),
+			},
+			//labels: []string{"ns", ""},
+			labels: []string{"ns", "ӕ"},
+			want: `label:<name:"namespace" value:"ns" > label:<name:"set" value:"\323\225" > gauge:<value:1 > `,
+		},
 	} {
 		metrics := cmetrics{c.field: c.metric}
-		ms := infoCollect(metrics, c.payload)
+		ms := infoCollect(metrics, c.payload, c.labels...)
 
 		if have, want := len(ms), 1; have != want {
 			t.Fatalf("have %d, want %d", have, want)
@@ -117,4 +150,4 @@ func TestInfoCollect(t *testing.T) {
 			t.Errorf("case %d: have %q, want %q", n, have, want)
 		}
 	}
-}
+}