From 5edec40d60a2d16dc946d4619193123667f17f54 Mon Sep 17 00:00:00 2001
From: Holger Hans Peter Freyther <holger@moiji-mobile.com>
Date: Mon, 23 Aug 2021 00:03:42 +0800
Subject: [PATCH] promtool: Speed up checking for duplicate rules

Trade space for speed. Convert all rules into our temporary struct, sort
and then iterate. This is a significant when having many rules.

Signed-off-by: Holger Hans Peter Freyther <holger@moiji-mobile.com>
---
 cmd/promtool/main.go      | 58 +++++++++++++++++++++++++++------------
 cmd/promtool/main_test.go |  2 ++
 2 files changed, 43 insertions(+), 17 deletions(-)

diff --git a/cmd/promtool/main.go b/cmd/promtool/main.go
index 1267631aa..242d90263 100644
--- a/cmd/promtool/main.go
+++ b/cmd/promtool/main.go
@@ -24,7 +24,7 @@ import (
 	"net/url"
 	"os"
 	"path/filepath"
-	"reflect"
+	"sort"
 	"strconv"
 	"strings"
 	"time"
@@ -48,6 +48,7 @@ import (
 	_ "github.com/prometheus/prometheus/discovery/install" // Register service discovery implementations.
 	"github.com/prometheus/prometheus/discovery/kubernetes"
 	"github.com/prometheus/prometheus/discovery/targetgroup"
+	"github.com/prometheus/prometheus/pkg/labels"
 	"github.com/prometheus/prometheus/pkg/rulefmt"
 	"github.com/prometheus/prometheus/promql"
 )
@@ -471,8 +472,8 @@ func checkRules(filename string) (int, []error) {
 		fmt.Printf("%d duplicate rule(s) found.\n", len(dRules))
 		for _, n := range dRules {
 			fmt.Printf("Metric: %s\nLabel(s):\n", n.metric)
-			for i, l := range n.label {
-				fmt.Printf("\t%s: %s\n", i, l)
+			for _, l := range n.label {
+				fmt.Printf("\t%s: %s\n", l.Name, l.Value)
 			}
 		}
 		fmt.Println("Might cause inconsistency while recording expressions.")
@@ -483,29 +484,52 @@ func checkRules(filename string) (int, []error) {
 
 type compareRuleType struct {
 	metric string
-	label  map[string]string
+	label  labels.Labels
+}
+
+type compareRuleTypes []compareRuleType
+
+func (c compareRuleTypes) Len() int           { return len(c) }
+func (c compareRuleTypes) Swap(i, j int)      { c[i], c[j] = c[j], c[i] }
+func (c compareRuleTypes) Less(i, j int) bool { return compare(c[i], c[j]) < 0 }
+
+func compare(a, b compareRuleType) int {
+	if res := strings.Compare(a.metric, b.metric); res != 0 {
+		return res
+	}
+
+	return labels.Compare(a.label, b.label)
 }
 
 func checkDuplicates(groups []rulefmt.RuleGroup) []compareRuleType {
 	var duplicates []compareRuleType
+	var rules compareRuleTypes
 
 	for _, group := range groups {
-		for index, rule := range group.Rules {
-			inst := compareRuleType{
+
+		for _, rule := range group.Rules {
+			rules = append(rules, compareRuleType{
 				metric: ruleMetric(rule),
-				label:  rule.Labels,
-			}
-			for i := 0; i < index; i++ {
-				t := compareRuleType{
-					metric: ruleMetric(group.Rules[i]),
-					label:  group.Rules[i].Labels,
-				}
-				if reflect.DeepEqual(t, inst) {
-					duplicates = append(duplicates, t)
-				}
-			}
+				label:  labels.FromMap(rule.Labels),
+			})
 		}
 	}
+	if len(rules) < 2 {
+		return duplicates
+	}
+	sort.Sort(rules)
+
+	last := rules[0]
+	for i := 1; i < len(rules); i++ {
+		if compare(last, rules[i]) == 0 {
+			// Don't add a duplicated rule multiple times.
+			if len(duplicates) == 0 || compare(last, duplicates[len(duplicates)-1]) != 0 {
+				duplicates = append(duplicates, rules[i])
+			}
+		}
+		last = rules[i]
+	}
+
 	return duplicates
 }
 
diff --git a/cmd/promtool/main_test.go b/cmd/promtool/main_test.go
index 6af9befee..87ac13e76 100644
--- a/cmd/promtool/main_test.go
+++ b/cmd/promtool/main_test.go
@@ -21,6 +21,7 @@ import (
 	"testing"
 	"time"
 
+	"github.com/prometheus/prometheus/pkg/labels"
 	"github.com/prometheus/prometheus/pkg/rulefmt"
 	"github.com/stretchr/testify/require"
 )
@@ -136,6 +137,7 @@ func TestCheckDuplicates(t *testing.T) {
 			expectedDups: []compareRuleType{
 				{
 					metric: "job:test:count_over_time1m",
+					label:  labels.New(),
 				},
 			},
 		},