Upgrade Thanos to latest master (#3363)

* Update Thanos to latest master.

Signed-off-by: Peter Štibraný <[email protected]>

* Added forgotten files
Signed-off-by: Peter Štibraný <[email protected]>

* Go mod tidy, vendor
Signed-off-by: Peter Štibraný <[email protected]>

* Set segment files when finishing block.
Signed-off-by: Peter Štibraný <[email protected]>

* Upgraded Thanos to latest master

Signed-off-by: Marco Pracucci <[email protected]>

* Updated CHANGELOG

Signed-off-by: Marco Pracucci <[email protected]>

* Rebased master

Signed-off-by: Marco Pracucci <[email protected]>

Co-authored-by: Peter Štibraný <[email protected]>

Author: pracucci

CHANGELOG.md

@@ -81,6 +81,7 @@
 * [ENHANCEMENT] Ruler: only load rules that belong to the ruler. Improves rules synching performances when ruler sharding is enabled. #3269
 * [ENHANCEMENT] Added `-<prefix>.redis.tls-insecure-skip-verify` flag. #3298
 * [ENHANCEMENT] Added `cortex_alertmanager_config_last_reload_successful_seconds` metric to show timestamp of last successful AM config reload. #3289
+* [ENHANCEMENT] Blocks storage: reduced number of bucket listing operations to list block content (applies to newly created blocks only). #3363
 * [BUGFIX] No-longer-needed ingester operations for queries triggered by queriers and rulers are now canceled. #3178
 * [BUGFIX] Ruler: directories in the configured `rules-path` will be removed on startup and shutdown in order to ensure they don't persist between runs. #3195
 * [BUGFIX] Handle hash-collisions in the query path. #3192

go.mod

@@ -10,7 +10,6 @@ require (
 	github.com/Masterminds/squirrel v0.0.0-20161115235646-20f192218cf5
 	github.com/NYTimes/gziphandler v1.1.1
 	github.com/alecthomas/units v0.0.0-20190924025748-f65c72e2690d
-	github.com/alicebob/gopher-json v0.0.0-20200520072559-a9ecdc9d1d3a // indirect
 	github.com/alicebob/miniredis v2.5.0+incompatible
 	github.com/armon/go-metrics v0.3.3
 	github.com/aws/aws-sdk-go v1.35.5
@@ -53,10 +52,9 @@ require (
 	github.com/sony/gobreaker v0.4.1
 	github.com/spf13/afero v1.2.2
 	github.com/stretchr/testify v1.6.1
-	github.com/thanos-io/thanos v0.13.1-0.20200923175059-57035bf8f843
+	github.com/thanos-io/thanos v0.13.1-0.20201019130456-f41940581d9a
 	github.com/uber/jaeger-client-go v2.25.0+incompatible
 	github.com/weaveworks/common v0.0.0-20200914083218-61ffdd448099
-	github.com/yuin/gopher-lua v0.0.0-20200816102855-ee81675732da // indirect
 	go.etcd.io/bbolt v1.3.5-0.20200615073812-232d8fc87f50
 	go.etcd.io/etcd v0.5.0-alpha.5.0.20200520232829-54ba9589114f
 	go.uber.org/atomic v1.7.0

go.sum

@@ -8,6 +8,7 @@ import (
 	"github.com/prometheus/prometheus/pkg/labels"
 	"github.com/prometheus/prometheus/storage"
 	"github.com/prometheus/prometheus/tsdb/chunkenc"
+	"github.com/thanos-io/thanos/pkg/store/labelpb"
 	"github.com/thanos-io/thanos/pkg/store/storepb"
 
 	"github.com/cortexproject/cortex/pkg/querier/series"
@@ -55,15 +56,15 @@ func (bqss *blockQuerierSeriesSet) Next() bool {
 		return false
 	}
 
-	currLabels := bqss.series[bqss.next].Labels
+	currLabels := labelpb.ZLabelsToPromLabels(bqss.series[bqss.next].Labels)
 	currChunks := bqss.series[bqss.next].Chunks
 
 	bqss.next++
 
 	// Merge chunks for current series. Chunks may come in multiple responses, but as soon
 	// as the response has chunks for a new series, we can stop searching. Series are sorted.
 	// See documentation for StoreClient.Series call for details.
-	for bqss.next < len(bqss.series) && storepb.CompareLabels(currLabels, bqss.series[bqss.next].Labels) == 0 {
+	for bqss.next < len(bqss.series) && labels.Compare(currLabels, labelpb.ZLabelsToPromLabels(bqss.series[bqss.next].Labels)) == 0 {
 		currChunks = append(currChunks, bqss.series[bqss.next].Chunks...)
 		bqss.next++
 	}
@@ -85,12 +86,12 @@ func (bqss *blockQuerierSeriesSet) Warnings() storage.Warnings {
 }
 
 // newBlockQuerierSeries makes a new blockQuerierSeries. Input labels must be already sorted by name.
-func newBlockQuerierSeries(lbls []storepb.Label, chunks []storepb.AggrChunk) *blockQuerierSeries {
+func newBlockQuerierSeries(lbls []labels.Label, chunks []storepb.AggrChunk) *blockQuerierSeries {
 	sort.Slice(chunks, func(i, j int) bool {
 		return chunks[i].MinTime < chunks[j].MinTime
 	})
 
-	return &blockQuerierSeries{labels: storepb.LabelsToPromLabelsUnsafe(lbls), chunks: chunks}
+	return &blockQuerierSeries{labels: lbls, chunks: chunks}
 }
 
 type blockQuerierSeries struct {

pkg/querier/block.go

@@ -14,6 +14,7 @@ import (
 	"github.com/prometheus/prometheus/tsdb/chunkenc"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
+	"github.com/thanos-io/thanos/pkg/store/labelpb"
 	"github.com/thanos-io/thanos/pkg/store/storepb"
 
 	"github.com/cortexproject/cortex/pkg/util"
@@ -39,7 +40,7 @@ func TestBlockQuerierSeries(t *testing.T) {
 		},
 		"should return series on success": {
 			series: &storepb.Series{
-				Labels: []storepb.Label{
+				Labels: []labelpb.ZLabel{
 					{Name: "foo", Value: "bar"},
 				},
 				Chunks: []storepb.AggrChunk{
@@ -56,7 +57,7 @@ func TestBlockQuerierSeries(t *testing.T) {
 		},
 		"should return error on failure while reading encoded chunk data": {
 			series: &storepb.Series{
-				Labels: []storepb.Label{{Name: "foo", Value: "bar"}},
+				Labels: []labelpb.ZLabel{{Name: "foo", Value: "bar"}},
 				Chunks: []storepb.AggrChunk{
 					{MinTime: minTimestamp.Unix() * 1000, MaxTime: maxTimestamp.Unix() * 1000, Raw: &storepb.Chunk{Type: storepb.Chunk_XOR, Data: []byte{0, 1}}},
 				},
@@ -70,7 +71,7 @@ func TestBlockQuerierSeries(t *testing.T) {
 		testData := testData
 
 		t.Run(testName, func(t *testing.T) {
-			series := newBlockQuerierSeries(testData.series.Labels, testData.series.Chunks)
+			series := newBlockQuerierSeries(labelpb.ZLabelsToPromLabels(testData.series.Labels), testData.series.Chunks)
 
 			assert.Equal(t, testData.expectedMetric, series.Labels())
 
@@ -119,23 +120,23 @@ func TestBlockQuerierSeriesSet(t *testing.T) {
 		series: []*storepb.Series{
 			// first, with one chunk.
 			{
-				Labels: mkLabels("__name__", "first", "a", "a"),
+				Labels: mkZLabels("__name__", "first", "a", "a"),
 				Chunks: []storepb.AggrChunk{
 					createAggrChunkWithSineSamples(now, now.Add(100*time.Second), 3*time.Millisecond), // ceil(100 / 0.003) samples (= 33334)
 				},
 			},
 
 			// continuation of previous series. Must have exact same labels.
 			{
-				Labels: mkLabels("__name__", "first", "a", "a"),
+				Labels: mkZLabels("__name__", "first", "a", "a"),
 				Chunks: []storepb.AggrChunk{
 					createAggrChunkWithSineSamples(now.Add(100*time.Second), now.Add(200*time.Second), 3*time.Millisecond), // ceil(100 / 0.003) samples more, 66668 in total
 				},
 			},
 
 			// second, with multiple chunks
 			{
-				Labels: mkLabels("__name__", "second"),
+				Labels: mkZLabels("__name__", "second"),
 				Chunks: []storepb.AggrChunk{
 					// unordered chunks
 					createAggrChunkWithSineSamples(now.Add(400*time.Second), now.Add(600*time.Second), 5*time.Millisecond), // 200 / 0.005 (= 40000 samples, = 120000 in total)
@@ -146,13 +147,13 @@ func TestBlockQuerierSeriesSet(t *testing.T) {
 
 			// overlapping
 			{
-				Labels: mkLabels("__name__", "overlapping"),
+				Labels: mkZLabels("__name__", "overlapping"),
 				Chunks: []storepb.AggrChunk{
 					createAggrChunkWithSineSamples(now, now.Add(10*time.Second), 5*time.Millisecond), // 10 / 0.005 = 2000 samples
 				},
 			},
 			{
-				Labels: mkLabels("__name__", "overlapping"),
+				Labels: mkZLabels("__name__", "overlapping"),
 				Chunks: []storepb.AggrChunk{
 					// 10 / 0.005 = 2000 samples, but first 1000 are overlapping with previous series, so this chunk only contributes 1000
 					createAggrChunkWithSineSamples(now.Add(5*time.Second), now.Add(15*time.Second), 5*time.Millisecond),
@@ -161,29 +162,29 @@ func TestBlockQuerierSeriesSet(t *testing.T) {
 
 			// overlapping 2. Chunks here come in wrong order.
 			{
-				Labels: mkLabels("__name__", "overlapping2"),
+				Labels: mkZLabels("__name__", "overlapping2"),
 				Chunks: []storepb.AggrChunk{
 					// entire range overlaps with the next chunk, so this chunks contributes 0 samples (it will be sorted as second)
 					createAggrChunkWithSineSamples(now.Add(3*time.Second), now.Add(7*time.Second), 5*time.Millisecond),
 				},
 			},
 			{
-				Labels: mkLabels("__name__", "overlapping2"),
+				Labels: mkZLabels("__name__", "overlapping2"),
 				Chunks: []storepb.AggrChunk{
 					// this chunk has completely overlaps previous chunk. Since its minTime is lower, it will be sorted as first.
 					createAggrChunkWithSineSamples(now, now.Add(10*time.Second), 5*time.Millisecond), // 10 / 0.005 = 2000 samples
 				},
 			},
 			{
-				Labels: mkLabels("__name__", "overlapping2"),
+				Labels: mkZLabels("__name__", "overlapping2"),
 				Chunks: []storepb.AggrChunk{
 					// no samples
 					createAggrChunkWithSineSamples(now, now, 5*time.Millisecond),
 				},
 			},
 
 			{
-				Labels: mkLabels("__name__", "overlapping2"),
+				Labels: mkZLabels("__name__", "overlapping2"),
 				Chunks: []storepb.AggrChunk{
 					// 2000 samples more (10 / 0.005)
 					createAggrChunkWithSineSamples(now.Add(20*time.Second), now.Add(30*time.Second), 5*time.Millisecond),
@@ -262,11 +263,11 @@ func createAggrChunk(minTime, maxTime int64, samples ...promql.Point) storepb.Ag
 	}
 }
 
-func mkLabels(s ...string) []storepb.Label {
-	result := []storepb.Label{}
+func mkZLabels(s ...string) []labelpb.ZLabel {
+	var result []labelpb.ZLabel
 
 	for i := 0; i+1 < len(s); i = i + 2 {
-		result = append(result, storepb.Label{
+		result = append(result, labelpb.ZLabel{
 			Name:  s[i],
 			Value: s[i+1],
 		})
@@ -275,6 +276,10 @@ func mkLabels(s ...string) []storepb.Label {
 	return result
 }
 
+func mkLabels(s ...string) []labels.Label {
+	return labelpb.ZLabelsToPromLabels(mkZLabels(s...))
+}
+
 func Benchmark_newBlockQuerierSeries(b *testing.B) {
 	lbls := mkLabels(
 		"__name__", "test",
@@ -308,7 +313,7 @@ func Benchmark_blockQuerierSeriesSet_iteration(b *testing.B) {
 	// Generate series.
 	series := make([]*storepb.Series, 0, numSeries)
 	for seriesID := 0; seriesID < numSeries; seriesID++ {
-		lbls := mkLabels("__name__", "test", "series_id", strconv.Itoa(seriesID))
+		lbls := mkZLabels("__name__", "test", "series_id", strconv.Itoa(seriesID))
 		chunks := make([]storepb.AggrChunk, 0, numChunksPerSeries)
 
 		// Create chunks with 1 sample per second.

pkg/querier/block_test.go

@@ -24,6 +24,7 @@ import (
 	"github.com/stretchr/testify/require"
 	"github.com/thanos-io/thanos/pkg/block/metadata"
 	"github.com/thanos-io/thanos/pkg/store/hintspb"
+	"github.com/thanos-io/thanos/pkg/store/labelpb"
 	"github.com/thanos-io/thanos/pkg/store/storepb"
 	"github.com/weaveworks/common/user"
 	"google.golang.org/grpc"
@@ -624,8 +625,8 @@ func TestBlocksStoreQuerier_SelectSortedShouldHonorQueryStoreAfter(t *testing.T)
 func TestBlocksStoreQuerier_PromQLExecution(t *testing.T) {
 	block1 := ulid.MustNew(1, nil)
 	block2 := ulid.MustNew(2, nil)
-	series1 := []storepb.Label{{Name: "__name__", Value: "metric_1"}}
-	series2 := []storepb.Label{{Name: "__name__", Value: "metric_2"}}
+	series1 := []labelpb.ZLabel{{Name: "__name__", Value: "metric_1"}}
+	series2 := []labelpb.ZLabel{{Name: "__name__", Value: "metric_2"}}
 
 	series1Samples := []promql.Point{
 		{T: 1589759955000, V: 1},
@@ -737,8 +738,8 @@ func TestBlocksStoreQuerier_PromQLExecution(t *testing.T) {
 	require.NoError(t, err)
 	require.Len(t, matrix, 2)
 
-	assert.Equal(t, storepb.LabelsToPromLabels(series1), matrix[0].Metric)
-	assert.Equal(t, storepb.LabelsToPromLabels(series2), matrix[1].Metric)
+	assert.Equal(t, labelpb.ZLabelsToPromLabels(series1), matrix[0].Metric)
+	assert.Equal(t, labelpb.ZLabelsToPromLabels(series2), matrix[1].Metric)
 	assert.Equal(t, series1Samples, matrix[0].Points)
 	assert.Equal(t, series2Samples, matrix[1].Points)
 }
@@ -840,7 +841,7 @@ func mockSeriesResponse(lbls labels.Labels, timeMillis int64, value float64) *st
 	return &storepb.SeriesResponse{
 		Result: &storepb.SeriesResponse_Series{
 			Series: &storepb.Series{
-				Labels: storepb.PromLabelsToLabels(lbls),
+				Labels: labelpb.ZLabelsFromPromLabels(lbls),
 				Chunks: []storepb.AggrChunk{
 					{MinTime: timeMillis, MaxTime: timeMillis, Raw: &storepb.Chunk{Type: storepb.Chunk_XOR, Data: chunkData}},
 				},

pkg/querier/blocks_store_queryable_test.go

@@ -18,6 +18,7 @@ import (
 	"github.com/stretchr/testify/mock"
 	"github.com/stretchr/testify/require"
 	"github.com/thanos-io/thanos/pkg/store"
+	"github.com/thanos-io/thanos/pkg/store/labelpb"
 	"github.com/thanos-io/thanos/pkg/store/storepb"
 	"github.com/weaveworks/common/logging"
 	"go.uber.org/atomic"
@@ -68,7 +69,7 @@ func TestBucketStores_InitialSync(t *testing.T) {
 		require.NoError(t, err)
 		assert.Empty(t, warnings)
 		require.Len(t, seriesSet, 1)
-		assert.Equal(t, []storepb.Label{{Name: labels.MetricName, Value: metricName}}, seriesSet[0].Labels)
+		assert.Equal(t, []labelpb.ZLabel{{Name: labels.MetricName, Value: metricName}}, seriesSet[0].Labels)
 	}
 
 	// Query series of another user.
@@ -146,7 +147,7 @@ func TestBucketStores_SyncBlocks(t *testing.T) {
 	require.NoError(t, err)
 	assert.Empty(t, warnings)
 	assert.Len(t, seriesSet, 1)
-	assert.Equal(t, []storepb.Label{{Name: labels.MetricName, Value: metricName}}, seriesSet[0].Labels)
+	assert.Equal(t, []labelpb.ZLabel{{Name: labels.MetricName, Value: metricName}}, seriesSet[0].Labels)
 
 	assert.NoError(t, testutil.GatherAndCompare(reg, strings.NewReader(`
 			# HELP cortex_bucket_store_blocks_loaded Number of currently loaded blocks.

pkg/storegateway/bucket_stores_test.go

@@ -25,6 +25,7 @@ import (
 	"github.com/stretchr/testify/require"
 	"github.com/thanos-io/thanos/pkg/block/metadata"
 	"github.com/thanos-io/thanos/pkg/extprom"
+	"github.com/thanos-io/thanos/pkg/store/labelpb"
 	"github.com/thanos-io/thanos/pkg/store/storepb"
 
 	"github.com/cortexproject/cortex/pkg/ring"
@@ -695,7 +696,7 @@ func TestStoreGateway_SeriesQueryingShouldRemoveExternalLabels(t *testing.T) {
 		actual := srv.SeriesSet[seriesID]
 
 		// Ensure Cortex external labels have been removed.
-		assert.Equal(t, []storepb.Label{{Name: "series_id", Value: strconv.Itoa(seriesID)}}, actual.Labels)
+		assert.Equal(t, []labelpb.ZLabel{{Name: "series_id", Value: strconv.Itoa(seriesID)}}, actual.Labels)
 
 		// Ensure samples have been correctly queried. The Thanos store also deduplicate samples
 		// in most cases, but it's not strictly required guaranteeing deduplication at this stage.

pkg/storegateway/gateway_test.go

@@ -19,6 +19,7 @@ import (
 	"github.com/prometheus/prometheus/tsdb/chunkenc"
 	"github.com/prometheus/prometheus/tsdb/chunks"
 	"github.com/prometheus/prometheus/tsdb/index"
+	"github.com/thanos-io/thanos/pkg/block"
 	"github.com/thanos-io/thanos/pkg/block/metadata"
 
 	"github.com/cortexproject/cortex/pkg/chunk"
@@ -204,8 +205,9 @@ func (d *tsdbBuilder) finishBlock(source string, labels map[string]string) (ulid
 		},
 
 		Thanos: metadata.Thanos{
-			Labels: labels,
-			Source: metadata.SourceType(source),
+			Labels:       labels,
+			Source:       metadata.SourceType(source),
+			SegmentFiles: block.GetSegmentFiles(d.tmpBlockDir),
 		},
 	}