promtool: allow setting multiple matchers to "promtool tsdb dump" command. (#13296)

Conditions are ANDed inside the same matcher but matchers are ORed

Including unit tests for "promtool tsdb dump".

Refactor some matchers scraping utils.

Signed-off-by: machine424 <ayoubmrini424@gmail.com>

cmd/promtool/main.go

@@ -236,7 +236,7 @@ func main() {
 	dumpPath := tsdbDumpCmd.Arg("db path", "Database path (default is "+defaultDBPath+").").Default(defaultDBPath).String()
 	dumpMinTime := tsdbDumpCmd.Flag("min-time", "Minimum timestamp to dump.").Default(strconv.FormatInt(math.MinInt64, 10)).Int64()
 	dumpMaxTime := tsdbDumpCmd.Flag("max-time", "Maximum timestamp to dump.").Default(strconv.FormatInt(math.MaxInt64, 10)).Int64()
-	dumpMatch := tsdbDumpCmd.Flag("match", "Series selector.").Default("{__name__=~'(?s:.*)'}").String()
+	dumpMatch := tsdbDumpCmd.Flag("match", "Series selector. Can be specified multiple times.").Default("{__name__=~'(?s:.*)'}").Strings()
 
 	importCmd := tsdbCmd.Command("create-blocks-from", "[Experimental] Import samples from input and produce TSDB blocks. Please refer to the storage docs for more details.")
 	importHumanReadable := importCmd.Flag("human-readable", "Print human readable values.").Short('r').Bool()

cmd/promtool/testdata/dump-test-1.prom

@@ -0,0 +1,15 @@
+{__name__="heavy_metric", foo="bar"} 5 0
+{__name__="heavy_metric", foo="bar"} 4 60000
+{__name__="heavy_metric", foo="bar"} 3 120000
+{__name__="heavy_metric", foo="bar"} 2 180000
+{__name__="heavy_metric", foo="bar"} 1 240000
+{__name__="heavy_metric", foo="foo"} 5 0
+{__name__="heavy_metric", foo="foo"} 4 60000
+{__name__="heavy_metric", foo="foo"} 3 120000
+{__name__="heavy_metric", foo="foo"} 2 180000
+{__name__="heavy_metric", foo="foo"} 1 240000
+{__name__="metric", baz="abc", foo="bar"} 1 0
+{__name__="metric", baz="abc", foo="bar"} 2 60000
+{__name__="metric", baz="abc", foo="bar"} 3 120000
+{__name__="metric", baz="abc", foo="bar"} 4 180000
+{__name__="metric", baz="abc", foo="bar"} 5 240000

cmd/promtool/testdata/dump-test-2.prom

@@ -0,0 +1,10 @@
+{__name__="heavy_metric", foo="foo"} 5 0
+{__name__="heavy_metric", foo="foo"} 4 60000
+{__name__="heavy_metric", foo="foo"} 3 120000
+{__name__="heavy_metric", foo="foo"} 2 180000
+{__name__="heavy_metric", foo="foo"} 1 240000
+{__name__="metric", baz="abc", foo="bar"} 1 0
+{__name__="metric", baz="abc", foo="bar"} 2 60000
+{__name__="metric", baz="abc", foo="bar"} 3 120000
+{__name__="metric", baz="abc", foo="bar"} 4 180000
+{__name__="metric", baz="abc", foo="bar"} 5 240000

cmd/promtool/testdata/dump-test-3.prom

@@ -0,0 +1,2 @@
+{__name__="metric", baz="abc", foo="bar"} 2 60000
+{__name__="metric", baz="abc", foo="bar"} 3 120000

cmd/promtool/tsdb.go

@@ -706,7 +706,7 @@ func analyzeCompaction(ctx context.Context, block tsdb.BlockReader, indexr tsdb.
 	return nil
 }
 
-func dumpSamples(ctx context.Context, path string, mint, maxt int64, match string) (err error) {
+func dumpSamples(ctx context.Context, path string, mint, maxt int64, match []string) (err error) {
 	db, err := tsdb.OpenDBReadOnly(path, nil)
 	if err != nil {
 		return err
@@ -720,11 +720,21 @@ func dumpSamples(ctx context.Context, path string, mint, maxt int64, match strin
 	}
 	defer q.Close()
 
-	matchers, err := parser.ParseMetricSelector(match)
+	matcherSets, err := parser.ParseMetricSelectors(match)
 	if err != nil {
 		return err
 	}
-	ss := q.Select(ctx, false, nil, matchers...)
+
+	var ss storage.SeriesSet
+	if len(matcherSets) > 1 {
+		var sets []storage.SeriesSet
+		for _, mset := range matcherSets {
+			sets = append(sets, q.Select(ctx, true, nil, mset...))
+		}
+		ss = storage.NewMergeSeriesSet(sets, storage.ChainedSeriesMerge)
+	} else {
+		ss = q.Select(ctx, false, nil, matcherSets[0]...)
+	}
 
 	for ss.Next() {
 		series := ss.At()

cmd/promtool/tsdb_test.go

@@ -14,9 +14,18 @@
 package main
 
 import (
+	"bytes"
+	"context"
+	"io"
+	"math"
+	"os"
+	"runtime"
+	"strings"
 	"testing"
 
 	"github.com/stretchr/testify/require"
+
+	"github.com/prometheus/prometheus/promql"
 )
 
 func TestGenerateBucket(t *testing.T) {
@@ -41,3 +50,101 @@ func TestGenerateBucket(t *testing.T) {
 		require.Equal(t, tc.step, step)
 	}
 }
+
+// getDumpedSamples dumps samples and returns them.
+func getDumpedSamples(t *testing.T, path string, mint, maxt int64, match []string) string {
+	t.Helper()
+
+	oldStdout := os.Stdout
+	r, w, _ := os.Pipe()
+	os.Stdout = w
+
+	err := dumpSamples(
+		context.Background(),
+		path,
+		mint,
+		maxt,
+		match,
+	)
+	require.NoError(t, err)
+
+	w.Close()
+	os.Stdout = oldStdout
+
+	var buf bytes.Buffer
+	io.Copy(&buf, r)
+	return buf.String()
+}
+
+func TestTSDBDump(t *testing.T) {
+	storage := promql.LoadedStorage(t, `
+		load 1m
+			metric{foo="bar", baz="abc"} 1 2 3 4 5
+			heavy_metric{foo="bar"} 5 4 3 2 1
+			heavy_metric{foo="foo"} 5 4 3 2 1
+	`)
+
+	tests := []struct {
+		name         string
+		mint         int64
+		maxt         int64
+		match        []string
+		expectedDump string
+	}{
+		{
+			name:         "default match",
+			mint:         math.MinInt64,
+			maxt:         math.MaxInt64,
+			match:        []string{"{__name__=~'(?s:.*)'}"},
+			expectedDump: "testdata/dump-test-1.prom",
+		},
+		{
+			name:         "same matcher twice",
+			mint:         math.MinInt64,
+			maxt:         math.MaxInt64,
+			match:        []string{"{foo=~'.+'}", "{foo=~'.+'}"},
+			expectedDump: "testdata/dump-test-1.prom",
+		},
+		{
+			name:         "no duplication",
+			mint:         math.MinInt64,
+			maxt:         math.MaxInt64,
+			match:        []string{"{__name__=~'(?s:.*)'}", "{baz='abc'}"},
+			expectedDump: "testdata/dump-test-1.prom",
+		},
+		{
+			name:         "well merged",
+			mint:         math.MinInt64,
+			maxt:         math.MaxInt64,
+			match:        []string{"{__name__='heavy_metric'}", "{baz='abc'}"},
+			expectedDump: "testdata/dump-test-1.prom",
+		},
+		{
+			name:         "multi matchers",
+			mint:         math.MinInt64,
+			maxt:         math.MaxInt64,
+			match:        []string{"{__name__='heavy_metric',foo='foo'}", "{__name__='metric'}"},
+			expectedDump: "testdata/dump-test-2.prom",
+		},
+		{
+			name:         "with reduced mint and maxt",
+			mint:         int64(60000),
+			maxt:         int64(120000),
+			match:        []string{"{__name__='metric'}"},
+			expectedDump: "testdata/dump-test-3.prom",
+		},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			dumpedMetrics := getDumpedSamples(t, storage.Dir(), tt.mint, tt.maxt, tt.match)
+			expectedMetrics, err := os.ReadFile(tt.expectedDump)
+			require.NoError(t, err)
+			if strings.Contains(runtime.GOOS, "windows") {
+				// We use "/n" while dumping on windows as well.
+				expectedMetrics = bytes.ReplaceAll(expectedMetrics, []byte("\r\n"), []byte("\n"))
+			}
+			// even though in case of one matcher samples are not sorted, the order in the cases above should stay the same.
+			require.Equal(t, string(expectedMetrics), dumpedMetrics)
+		})
+	}
+}

docs/command-line/promtool.md

@@ -567,7 +567,7 @@ Dump samples from a TSDB.
 | --- | --- | --- |
 | <code class="text-nowrap">--min-time</code> | Minimum timestamp to dump. | `-9223372036854775808` |
 | <code class="text-nowrap">--max-time</code> | Maximum timestamp to dump. | `9223372036854775807` |
-| <code class="text-nowrap">--match</code> | Series selector. | `{__name__=~'(?s:.*)'}` |
+| <code class="text-nowrap">--match</code> | Series selector. Can be specified multiple times. | `{__name__=~'(?s:.*)'}` |
 
 
 

promql/parser/parse.go

@@ -208,6 +208,20 @@ func ParseMetricSelector(input string) (m []*labels.Matcher, err error) {
 	return m, err
 }
 
+// ParseMetricSelectors parses a list of provided textual metric selectors into lists of
+// label matchers.
+func ParseMetricSelectors(matchers []string) (m [][]*labels.Matcher, err error) {
+	var matcherSets [][]*labels.Matcher
+	for _, s := range matchers {
+		matchers, err := ParseMetricSelector(s)
+		if err != nil {
+			return nil, err
+		}
+		matcherSets = append(matcherSets, matchers)
+	}
+	return matcherSets, nil
+}
+
 // SequenceValue is an omittable value in a sequence of time series values.
 type SequenceValue struct {
 	Value     float64

web/api/v1/api.go

@@ -1848,13 +1848,9 @@ func parseDuration(s string) (time.Duration, error) {
 }
 
 func parseMatchersParam(matchers []string) ([][]*labels.Matcher, error) {
-	var matcherSets [][]*labels.Matcher
-	for _, s := range matchers {
-		matchers, err := parser.ParseMetricSelector(s)
-		if err != nil {
-			return nil, err
-		}
-		matcherSets = append(matcherSets, matchers)
+	matcherSets, err := parser.ParseMetricSelectors(matchers)
+	if err != nil {
+		return nil, err
 	}
 
 OUTER:

web/federate.go

@@ -65,14 +65,10 @@ func (h *Handler) federation(w http.ResponseWriter, req *http.Request) {
 		return
 	}
 
-	var matcherSets [][]*labels.Matcher
-	for _, s := range req.Form["match[]"] {
-		matchers, err := parser.ParseMetricSelector(s)
-		if err != nil {
-			http.Error(w, err.Error(), http.StatusBadRequest)
-			return
-		}
-		matcherSets = append(matcherSets, matchers)
+	matcherSets, err := parser.ParseMetricSelectors(req.Form["match[]"])
+	if err != nil {
+		http.Error(w, err.Error(), http.StatusBadRequest)
+		return
 	}
 
 	var (