model/textparse: parsers take a labels SymbolTable

This allows strings to be interned to save memory.

Signed-off-by: Bryan Boreham <bjboreham@gmail.com>

cmd/promtool/backfill.go

@@ -127,7 +127,8 @@ func createBlocks(input []byte, mint, maxt, maxBlockDuration int64, maxSamplesIn
 
 			ctx := context.Background()
 			app := w.Appender(ctx)
-			p := textparse.NewOpenMetricsParser(input)
+			symbolTable := labels.NewSymbolTable() // One table per block means it won't grow too large.
+			p := textparse.NewOpenMetricsParser(input, symbolTable)
 			samplesCount := 0
 			for {
 				e, err := p.Next()
@@ -216,7 +217,7 @@ func createBlocks(input []byte, mint, maxt, maxBlockDuration int64, maxSamplesIn
 }
 
 func backfill(maxSamplesInAppender int, input []byte, outputDir string, humanReadable, quiet bool, maxBlockDuration time.Duration) (err error) {
-	p := textparse.NewOpenMetricsParser(input)
+	p := textparse.NewOpenMetricsParser(input, nil) // Don't need a SymbolTable to get max and min timestamps.
 	maxt, mint, err := getMinAndMaxTimestamps(p)
 	if err != nil {
 		return fmt.Errorf("getting min and max timestamp: %w", err)

model/textparse/interface.go

@@ -80,22 +80,22 @@ type Parser interface {
 //
 // This function always returns a valid parser, but might additionally
 // return an error if the content type cannot be parsed.
-func New(b []byte, contentType string, parseClassicHistograms bool) (Parser, error) {
+func New(b []byte, contentType string, parseClassicHistograms bool, st *labels.SymbolTable) (Parser, error) {
 	if contentType == "" {
-		return NewPromParser(b), nil
+		return NewPromParser(b, st), nil
 	}
 
 	mediaType, _, err := mime.ParseMediaType(contentType)
 	if err != nil {
-		return NewPromParser(b), err
+		return NewPromParser(b, st), err
 	}
 	switch mediaType {
 	case "application/openmetrics-text":
-		return NewOpenMetricsParser(b), nil
+		return NewOpenMetricsParser(b, st), nil
 	case "application/vnd.google.protobuf":
-		return NewProtobufParser(b, parseClassicHistograms), nil
+		return NewProtobufParser(b, parseClassicHistograms, st), nil
 	default:
-		return NewPromParser(b), nil
+		return NewPromParser(b, st), nil
 	}
 }
 

model/textparse/interface_test.go

@@ -17,6 +17,8 @@ import (
 	"testing"
 
 	"github.com/stretchr/testify/require"
+
+	"github.com/prometheus/prometheus/model/labels"
 )
 
 func TestNewParser(t *testing.T) {
@@ -91,7 +93,7 @@ func TestNewParser(t *testing.T) {
 			tt := tt // Copy to local variable before going parallel.
 			t.Parallel()
 
-			p, err := New([]byte{}, tt.contentType, false)
+			p, err := New([]byte{}, tt.contentType, false, labels.NewSymbolTable())
 			tt.validateParser(t, p)
 			if tt.err == "" {
 				require.NoError(t, err)

model/textparse/openmetricsparse.go

@@ -97,8 +97,11 @@ type OpenMetricsParser struct {
 }
 
 // NewOpenMetricsParser returns a new parser of the byte slice.
-func NewOpenMetricsParser(b []byte) Parser {
-	return &OpenMetricsParser{l: &openMetricsLexer{b: b}}
+func NewOpenMetricsParser(b []byte, st *labels.SymbolTable) Parser {
+	return &OpenMetricsParser{
+		l:       &openMetricsLexer{b: b},
+		builder: labels.NewScratchBuilderWithSymbolTable(st, 16),
+	}
 }
 
 // Series returns the bytes of the series, the timestamp if set, and the value

model/textparse/openmetricsparse_test.go

@@ -247,7 +247,7 @@ foo_total 17.0 1520879607.789 # {id="counter-test"} 5`
 		},
 	}
 
-	p := NewOpenMetricsParser([]byte(input))
+	p := NewOpenMetricsParser([]byte(input), labels.NewSymbolTable())
 	i := 0
 
 	var res labels.Labels
@@ -378,7 +378,7 @@ choices}`, "strange©™\n'quoted' \"name\"", "6"),
 		},
 	}
 
-	p := NewOpenMetricsParser([]byte(input))
+	p := NewOpenMetricsParser([]byte(input), labels.NewSymbolTable())
 	i := 0
 
 	var res labels.Labels
@@ -727,7 +727,7 @@ func TestOpenMetricsParseErrors(t *testing.T) {
 	}
 
 	for i, c := range cases {
-		p := NewOpenMetricsParser([]byte(c.input))
+		p := NewOpenMetricsParser([]byte(c.input), labels.NewSymbolTable())
 		var err error
 		for err == nil {
 			_, err = p.Next()
@@ -792,7 +792,7 @@ func TestOMNullByteHandling(t *testing.T) {
 	}
 
 	for i, c := range cases {
-		p := NewOpenMetricsParser([]byte(c.input))
+		p := NewOpenMetricsParser([]byte(c.input), labels.NewSymbolTable())
 		var err error
 		for err == nil {
 			_, err = p.Next()

model/textparse/promparse.go

@@ -166,8 +166,11 @@ type PromParser struct {
 }
 
 // NewPromParser returns a new parser of the byte slice.
-func NewPromParser(b []byte) Parser {
-	return &PromParser{l: &promlexer{b: append(b, '\n')}}
+func NewPromParser(b []byte, st *labels.SymbolTable) Parser {
+	return &PromParser{
+		l:       &promlexer{b: append(b, '\n')},
+		builder: labels.NewScratchBuilderWithSymbolTable(st, 16),
+	}
 }
 
 // Series returns the bytes of the series, the timestamp if set, and the value

model/textparse/promparse_test.go

@@ -178,7 +178,7 @@ testmetric{label="\"bar\""} 1`
 		},
 	}
 
-	p := NewPromParser([]byte(input))
+	p := NewPromParser([]byte(input), labels.NewSymbolTable())
 	i := 0
 
 	var res labels.Labels
@@ -304,7 +304,7 @@ choices}`, "strange©™\n'quoted' \"name\"", "6"),
 		},
 	}
 
-	p := NewPromParser([]byte(input))
+	p := NewPromParser([]byte(input), labels.NewSymbolTable())
 	i := 0
 
 	var res labels.Labels
@@ -422,7 +422,7 @@ func TestPromParseErrors(t *testing.T) {
 	}
 
 	for i, c := range cases {
-		p := NewPromParser([]byte(c.input))
+		p := NewPromParser([]byte(c.input), labels.NewSymbolTable())
 		var err error
 		for err == nil {
 			_, err = p.Next()
@@ -476,7 +476,7 @@ func TestPromNullByteHandling(t *testing.T) {
 	}
 
 	for i, c := range cases {
-		p := NewPromParser([]byte(c.input))
+		p := NewPromParser([]byte(c.input), labels.NewSymbolTable())
 		var err error
 		for err == nil {
 			_, err = p.Next()
@@ -497,7 +497,7 @@ const (
 )
 
 func BenchmarkParse(b *testing.B) {
-	for parserName, parser := range map[string]func([]byte) Parser{
+	for parserName, parser := range map[string]func([]byte, *labels.SymbolTable) Parser{
 		"prometheus":  NewPromParser,
 		"openmetrics": NewOpenMetricsParser,
 	} {
@@ -516,8 +516,9 @@ func BenchmarkParse(b *testing.B) {
 				b.ReportAllocs()
 				b.ResetTimer()
 
+				st := labels.NewSymbolTable()
 				for i := 0; i < b.N; i += promtestdataSampleCount {
-					p := parser(buf)
+					p := parser(buf, st)
 
 				Outer:
 					for i < b.N {
@@ -544,8 +545,9 @@ func BenchmarkParse(b *testing.B) {
 				b.ReportAllocs()
 				b.ResetTimer()
 
+				st := labels.NewSymbolTable()
 				for i := 0; i < b.N; i += promtestdataSampleCount {
-					p := parser(buf)
+					p := parser(buf, st)
 
 				Outer:
 					for i < b.N {
@@ -577,8 +579,9 @@ func BenchmarkParse(b *testing.B) {
 				b.ReportAllocs()
 				b.ResetTimer()
 
+				st := labels.NewSymbolTable()
 				for i := 0; i < b.N; i += promtestdataSampleCount {
-					p := parser(buf)
+					p := parser(buf, st)
 
 				Outer:
 					for i < b.N {

model/textparse/protobufparse.go

@@ -80,13 +80,14 @@ type ProtobufParser struct {
 }
 
 // NewProtobufParser returns a parser for the payload in the byte slice.
-func NewProtobufParser(b []byte, parseClassicHistograms bool) Parser {
+func NewProtobufParser(b []byte, parseClassicHistograms bool, st *labels.SymbolTable) Parser {
 	return &ProtobufParser{
 		in:                     b,
 		state:                  EntryInvalid,
 		mf:                     &dto.MetricFamily{},
 		metricBytes:            &bytes.Buffer{},
 		parseClassicHistograms: parseClassicHistograms,
+		builder:                labels.NewScratchBuilderWithSymbolTable(st, 16),
 	}
 }
 

model/textparse/protobufparse_test.go

@@ -743,7 +743,7 @@ func TestProtobufParse(t *testing.T) {
 	}{
 		{
 			name:   "ignore classic buckets of native histograms",
-			parser: NewProtobufParser(inputBuf.Bytes(), false),
+			parser: NewProtobufParser(inputBuf.Bytes(), false, labels.NewSymbolTable()),
 			expected: []parseResult{
 				{
 					m:    "go_build_info",
@@ -1280,7 +1280,7 @@ func TestProtobufParse(t *testing.T) {
 		},
 		{
 			name:   "parse classic and native buckets",
-			parser: NewProtobufParser(inputBuf.Bytes(), true),
+			parser: NewProtobufParser(inputBuf.Bytes(), true, labels.NewSymbolTable()),
 			expected: []parseResult{
 				{ // 0
 					m:    "go_build_info",

web/federate_test.go

@@ -391,7 +391,7 @@ func TestFederationWithNativeHistograms(t *testing.T) {
 	body, err := io.ReadAll(res.Body)
 	require.NoError(t, err)
 
-	p := textparse.NewProtobufParser(body, false)
+	p := textparse.NewProtobufParser(body, false, labels.NewSymbolTable())
 	var actVec promql.Vector
 	metricFamilies := 0
 	l := labels.Labels{}