Add search block headers (#943)

* Add search block headers Signed-off-by: Martin Disibio <[email protected]> * nolint Signed-off-by: Martin Disibio <[email protected]> * Fix buffer usage in SearchBlockIterator to prevent data corruption Signed-off-by: Martin Disibio <[email protected]> * Reorganize search iterator and combiner Signed-off-by: Martin Disibio <[email protected]> * changelog Signed-off-by: Martin Disibio <[email protected]> * no nolint Signed-off-by: Martin Disibio <[email protected]> * Add skipped blocks metric to response Signed-off-by: Martin Disibio <[email protected]>
grafana · Sep 15, 2021 · 0e7e7aa · 0e7e7aa
1 parent 78d8107
commit 0e7e7aa
Show file tree

Hide file tree

Showing 16 changed files with 529 additions and 186 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -35,6 +35,7 @@
 * [ENHANCEMENT] Implement trace comparison in Vulture [#904](https://github.com/grafana/tempo/pull/904) (@zalegrala)
 * [ENHANCEMENT] Dedupe search records while replaying WAL [#940](https://github.com/grafana/tempo/pull/940) (@annanay25)
 * [ENHANCEMENT] Add status endpoint to list the available endpoints [#938](https://github.com/grafana/tempo/pull/938) (@zalegrala)
+* [ENHANCEMENT] Add search block headers [#943](https://github.com/grafana/tempo/pull/943) (@mdisibio)
 * [CHANGE] Renamed CLI flag from `--storage.trace.maintenance-cycle` to `--storage.trace.blocklist_poll`. This is a **breaking change**  [#897](https://github.com/grafana/tempo/pull/897) (@mritunjaysharma394)
 * [CHANGE] update jsonnet alerts and recording rules to use `job_selectors` and `cluster_selectors` for configurable unique identifier labels [#935](https://github.com/grafana/tempo/pull/935) (@kevinschoonover)
 * [CHANGE] Modify generated tag keys in Vulture for easier filtering [#934](https://github.com/grafana/tempo/pull/934) (@zalegrala)

diff --git a/modules/ingester/instance_search.go b/modules/ingester/instance_search.go
@@ -61,6 +61,7 @@ func (i *instance) Search(ctx context.Context, req *tempopb.SearchRequest) (*tem
 			InspectedTraces: sr.TracesInspected(),
 			InspectedBytes:  sr.BytesInspected(),
 			InspectedBlocks: sr.BlocksInspected(),
+			SkippedBlocks:   sr.BlocksSkipped(),
 		},
 	}, nil
 }

diff --git a/modules/ingester/instance_search_test.go b/modules/ingester/instance_search_test.go
@@ -17,6 +17,7 @@ import (
 	"github.com/grafana/tempo/pkg/tempopb"
 	"github.com/grafana/tempo/pkg/util"
 	"github.com/grafana/tempo/pkg/util/test"
+	"github.com/grafana/tempo/tempodb/search"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
 )
@@ -357,7 +358,8 @@ func TestInstanceSearchMetrics(t *testing.T) {
 
 	search := func() *tempopb.SearchMetrics {
 		sr, err := i.Search(context.Background(), &tempopb.SearchRequest{
-			Tags: map[string]string{"nomatch": "nomatch"},
+			// Exhaustive search
+			Tags: map[string]string{search.SecretExhaustiveSearchTag: "!"},
 		})
 		require.NoError(t, err)
 		return sr.Metrics

diff --git a/modules/querier/querier.go b/modules/querier/querier.go
@@ -375,6 +375,7 @@ func (q *Querier) postProcessSearchResults(req *tempopb.SearchRequest, rr []resp
 			response.Metrics.InspectedBytes += sr.Metrics.InspectedBytes
 			response.Metrics.InspectedTraces += sr.Metrics.InspectedTraces
 			response.Metrics.InspectedBlocks += sr.Metrics.InspectedBlocks
+			response.Metrics.SkippedBlocks += sr.Metrics.SkippedBlocks
 		}
 	}
 

diff --git a/pkg/tempofb/SearchBlockHeader.go b/pkg/tempofb/SearchBlockHeader.go
diff --git a/pkg/tempofb/SearchBlockHeader_util.go b/pkg/tempofb/SearchBlockHeader_util.go
@@ -0,0 +1,56 @@
+package tempofb
+
+import flatbuffers "github.com/google/flatbuffers/go"
+
+type SearchBlockHeaderBuilder struct {
+	Tags   SearchDataMap
+	MinDur uint64
+	MaxDur uint64
+}
+
+func NewSearchBlockHeaderBuilder() *SearchBlockHeaderBuilder {
+	return &SearchBlockHeaderBuilder{
+		Tags: SearchDataMap{},
+	}
+}
+
+func (s *SearchBlockHeaderBuilder) AddEntry(e *SearchEntry) {
+
+	kv := &KeyValues{} //buffer
+
+	// Record all unique keyvalues
+	for i, ii := 0, e.TagsLength(); i < ii; i++ {
+		e.Tags(kv, i)
+		for j, jj := 0, kv.ValueLength(); j < jj; j++ {
+			s.AddTag(string(kv.Key()), string(kv.Value(j)))
+		}
+	}
+
+	// Record min/max durations
+	dur := e.EndTimeUnixNano() - e.StartTimeUnixNano()
+	if s.MinDur == 0 || dur < s.MinDur {
+		s.MinDur = dur
+	}
+	if dur > s.MaxDur {
+		s.MaxDur = dur
+	}
+}
+
+// AddTag adds the unique tag name and value to the search data. No effect if the pair is already present.
+func (s *SearchBlockHeaderBuilder) AddTag(k string, v string) {
+	s.Tags.Add(k, v)
+}
+
+func (s *SearchBlockHeaderBuilder) ToBytes() []byte {
+	b := flatbuffers.NewBuilder(1024)
+
+	tags := s.Tags.WriteToBuilder(b)
+
+	SearchBlockHeaderStart(b)
+	SearchBlockHeaderAddMinDurationNanos(b, s.MinDur)
+	SearchBlockHeaderAddMaxDurationNanos(b, s.MaxDur)
+	SearchBlockHeaderAddTags(b, tags)
+	offset := SearchBlockHeaderEnd(b)
+	b.Finish(offset)
+	return b.FinishedBytes()
+}
diff --git a/pkg/tempofb/tempo.fbs b/pkg/tempofb/tempo.fbs
@@ -22,4 +22,16 @@ table SearchPage {
 
     // Trace entries
     entries : [SearchEntry];
+}
+
+table SearchBlockHeader {
+    // This is a rollup of all distinct tags/values in the
+    // block for quick elimination.
+    tags : [KeyValues];
+
+    // Smallest trace duration in the block
+    min_duration_nanos: uint64;
+
+    // Largest trace duration in the block
+    max_duration_nanos: uint64;
 }