fix: replace unprintable and invalid characters in errors (#23387)

Replace unprintable and invalid characters with '?'
in logged errors.  Truncate consecutive runs of them to
only 3 repeats of '?'

closes #23386

tsdb/shard.go

@@ -15,6 +15,7 @@ import (
 	"sync"
 	"sync/atomic"
 	"time"
+	"unicode"
 	"unsafe"
 
 	"github.com/influxdata/influxdb/models"
@@ -601,7 +602,7 @@ func (s *Shard) validateSeriesAndFields(points []models.Point, tracker StatsTrac
 		if validateKeys && !models.ValidKeyTokens(string(p.Name()), tags) {
 			dropped++
 			if reason == "" {
-				reason = fmt.Sprintf("key contains invalid unicode: \"%s\"", string(p.Key()))
+				reason = fmt.Sprintf("key contains invalid unicode: %q", makePrintable(string(p.Key())))
 			}
 			continue
 		}
@@ -723,6 +724,28 @@ func (s *Shard) validateSeriesAndFields(points []models.Point, tracker StatsTrac
 	return points[:j], fieldsToCreate, err
 }
 
+const unPrintReplRune = '?'
+const unPrintMaxReplRune = 3
+
+// makePrintable - replace invalid and non-printable unicode characters with a few '?' runes
+func makePrintable(s string) string {
+	b := strings.Builder{}
+	b.Grow(len(s))
+	c := 0
+	for _, r := range strings.ToValidUTF8(s, string(unicode.ReplacementChar)) {
+		if !unicode.IsPrint(r) || r == unicode.ReplacementChar {
+			if c < unPrintMaxReplRune {
+				b.WriteRune(unPrintReplRune)
+			}
+			c++
+		} else {
+			b.WriteRune(r)
+			c = 0
+		}
+	}
+	return b.String()
+}
+
 func (s *Shard) createFieldsAndMeasurements(fieldsToCreate []*FieldCreate) error {
 	if len(fieldsToCreate) == 0 {
 		return nil

tsdb/shard_internal_test.go

@@ -16,8 +16,40 @@ import (
 	"github.com/influxdata/influxdb/logger"
 	"github.com/influxdata/influxdb/models"
 	"github.com/influxdata/influxql"
+	"github.com/stretchr/testify/require"
 )
 
+func TestShard_ErrorPrinting(t *testing.T) {
+
+	tests := []struct {
+		nSeq int
+		raw  string
+	}{
+		{1, string([]byte{'b', 'e', 'n', 't', 'e', 's', 't', '\t', '\n'})},
+		{1, string([]byte{'b', 'e', 'n', 't', 'e', 's', 0, 0, 0xFE, 0, 0xFE, 't'})},
+		{2, string([]byte{0, 0, 0, 0, 0xFE, '\t', '\n', '\t', 'b', 'e', 'n', 't', 'e', 's', 't', 0, 0, 0, 0, 0xFE, '\t', '\n', '\t', '\t', '\t'})},
+	}
+
+	for i, _ := range tests {
+		f := makePrintable(tests[i].raw)
+		require.True(t, models.ValidKeyToken(f))
+		c := 0
+		nSeq := 0
+		for _, r := range f {
+			if r == unPrintReplRune {
+				c++
+				if c == 1 {
+					nSeq++
+				}
+				require.LessOrEqual(t, c, unPrintMaxReplRune, "too many repeated %c", unPrintReplRune)
+			} else {
+				c = 0
+			}
+		}
+		require.Equalf(t, tests[i].nSeq, nSeq, "wrong number of elided sequences of replacement characters")
+	}
+}
+
 func TestShard_MapType(t *testing.T) {
 	var sh *TempShard