tempo-cli analyse blocks skips compacted blocks and updates block list (

grafana#3183)

CHANGELOG.md

@@ -1,6 +1,7 @@
 ## main / unreleased
 
-+ [BUGFIX] Change exit code if config is successfully verified [#3174](https://github.com/grafana/tempo/pull/3174) (@am3o @agrib-01)
+* [BUGFIX] Change exit code if config is successfully verified [#3174](https://github.com/grafana/tempo/pull/3174) (@am3o @agrib-01)
+* [BUGFIX] The tempo-cli analyse blocks command no longer fails on compacted blocks [#3183](https://github.com/grafana/tempo/pull/3183) (@stoewer)
 * [ENHANCEMENT] Introduced `AttributePolicyMatch` & `IntrinsicPolicyMatch` structures to match span attributes based on strongly typed values & precompiled regexp [#3025](https://github.com/grafana/tempo/pull/3025) (@andriusluk)
 * [CHANGE] TraceQL/Structural operators performance improvement. [#3088](https://github.com/grafana/tempo/pull/3088) (@joe-elliott)
 * [CHANGE] Merge the processors overrides set through runtime overrides and user-configurable overrides [#3125](https://github.com/grafana/tempo/pull/3125) (@kvrhdn)

cmd/tempo-cli/cmd-analyse-block.go

@@ -104,6 +104,9 @@ func (cmd *analyseBlockCmd) Run(ctx *globalOptions) error {
 
 	blockSum, err := processBlock(r, c, cmd.TenantID, cmd.BlockID, time.Hour, 0)
 	if err != nil {
+		if errors.Is(err, backend.ErrDoesNotExist) {
+			return fmt.Errorf("unable to analyze block: block has no block.meta because it was compacted")
+		}
 		return err
 	}
 
@@ -118,15 +121,9 @@ func processBlock(r backend.Reader, _ backend.Compactor, tenantID, blockID strin
 	id := uuid.MustParse(blockID)
 
 	meta, err := r.BlockMeta(context.TODO(), id, tenantID)
-	if err != nil && !errors.Is(err, backend.ErrDoesNotExist) {
+	if err != nil {
 		return nil, err
 	}
-
-	if meta == nil {
-		fmt.Println("Unable to load any meta for block", blockID)
-		return nil, nil
-	}
-
 	if meta.CompactionLevel < minCompactionLvl {
 		return nil, nil
 	}
@@ -344,7 +341,7 @@ func printSummary(scope string, max int, summary genericAttrSummary) error {
 	return w.Flush()
 }
 
-func printDedicatedColumnOverridesJsonnet(spanSummary genericAttrSummary, resourceSummary genericAttrSummary) {
+func printDedicatedColumnOverridesJsonnet(spanSummary, resourceSummary genericAttrSummary) {
 	fmt.Println("")
 	fmt.Printf("parquet_dedicated_columns: [\n")
 

cmd/tempo-cli/cmd-analyse-blocks.go

@@ -2,7 +2,12 @@ package main
 
 import (
 	"context"
+	"errors"
 	"time"
+
+	"github.com/google/uuid"
+
+	"github.com/grafana/tempo/tempodb/backend"
 )
 
 type analyseBlocksCmd struct {
@@ -26,17 +31,30 @@ func (cmd *analyseBlocksCmd) Run(ctx *globalOptions) error {
 		return err
 	}
 
-	processedBlocks := 0
+	processedBlocks := map[uuid.UUID]struct{}{}
 	topSpanAttrs, topResourceAttrs := make(map[string]uint64), make(map[string]uint64)
 	totalSpanBytes, totalResourceBytes := uint64(0), uint64(0)
-	for _, block := range blocks {
-		if processedBlocks >= cmd.MaxBlocks {
-			break
+
+	for i := 0; i < len(blocks) && len(processedBlocks) < cmd.MaxBlocks; i++ {
+		block := blocks[i]
+		if _, ok := processedBlocks[block]; ok {
+			continue
 		}
 
 		blockSum, err := processBlock(r, c, cmd.TenantID, block.String(), time.Hour, uint8(cmd.MinCompactionLevel))
 		if err != nil {
-			return err
+			if !errors.Is(err, backend.ErrDoesNotExist) {
+				return err
+			}
+
+			// the block was already compacted and blocks might be outdated: refreshing blocks
+			blocks, _, err = r.Blocks(context.Background(), cmd.TenantID)
+			if err != nil {
+				return err
+			}
+			i = -1
+
+			continue
 		}
 
 		if blockSum == nil {
@@ -53,8 +71,9 @@ func (cmd *analyseBlocksCmd) Run(ctx *globalOptions) error {
 		}
 		totalResourceBytes += blockSum.resourceSummary.totalBytes
 
-		processedBlocks++
+		processedBlocks[block] = struct{}{}
 	}
+
 	// Get top N attributes from map
 	return (&blockSummary{
 		spanSummary: genericAttrSummary{