fix: validate scrape_config job name

Author: fgouteroux

clients/pkg/promtail/promtail.go

@@ -18,6 +18,7 @@ import (
 	"github.com/grafana/loki/v3/clients/pkg/promtail/api"
 	"github.com/grafana/loki/v3/clients/pkg/promtail/client"
 	"github.com/grafana/loki/v3/clients/pkg/promtail/config"
+	"github.com/grafana/loki/v3/clients/pkg/promtail/scrapeconfig"
 	"github.com/grafana/loki/v3/clients/pkg/promtail/server"
 	"github.com/grafana/loki/v3/clients/pkg/promtail/targets"
 	"github.com/grafana/loki/v3/clients/pkg/promtail/targets/target"
@@ -139,10 +140,16 @@ func (p *Promtail) reloadConfig(cfg *config.Config) error {
 	}
 
 	cfg.Setup(p.logger)
+	var err error
+	err = scrapeconfig.ValidateJobName(cfg.ScrapeConfig)
+	if err != nil {
+		return err
+	}
+
 	if cfg.LimitsConfig.ReadlineRateEnabled {
 		stages.SetReadLineRateLimiter(cfg.LimitsConfig.ReadlineRate, cfg.LimitsConfig.ReadlineBurst, cfg.LimitsConfig.ReadlineRateDrop)
 	}
-	var err error
+
 	// entryHandlers contains all sinks were scraped log entries should get to
 	var entryHandlers = []api.EntryHandler{}
 

clients/pkg/promtail/scrapeconfig/scrapeconfig.go

@@ -1,6 +1,7 @@
 package scrapeconfig
 
 import (
+	"errors"
 	"fmt"
 	"reflect"
 	"time"
@@ -26,6 +27,7 @@ import (
 	"github.com/prometheus/prometheus/discovery/triton"
 	"github.com/prometheus/prometheus/discovery/zookeeper"
 	"github.com/prometheus/prometheus/model/relabel"
+	"github.com/prometheus/prometheus/util/strutil"
 
 	"github.com/grafana/loki/v3/clients/pkg/logentry/stages"
 	"github.com/grafana/loki/v3/clients/pkg/promtail/discovery/consulagent"
@@ -484,3 +486,22 @@ func (c *Config) UnmarshalYAML(unmarshal func(interface{}) error) error {
 
 	return nil
 }
+
+func ValidateJobName(scrapeConfigs []Config) error {
+	jobNames := map[string]struct{}{}
+	for i, cfg := range scrapeConfigs {
+		if cfg.JobName == "" {
+			return errors.New("`job_name` must be defined for the scrape_config with a " +
+				"unique name, " +
+				"at least one scrape_config has no `job_name` defined")
+		}
+		if _, ok := jobNames[cfg.JobName]; ok {
+			return fmt.Errorf("`job_name` must be unique for each scrape_config, "+
+				"a duplicate `job_name` of %s was found", cfg.JobName)
+		}
+		jobNames[cfg.JobName] = struct{}{}
+
+		scrapeConfigs[i].JobName = strutil.SanitizeLabelName(cfg.JobName)
+	}
+	return nil
+}

clients/pkg/promtail/scrapeconfig/scrapeconfig_test.go

@@ -142,3 +142,69 @@ func TestLoadConfig(t *testing.T) {
 
 	require.NotZero(t, len(config.PipelineStages))
 }
+
+func Test_validateJobName(t *testing.T) {
+	tests := []struct {
+		name    string
+		configs []Config
+		// Only validated against the first job in the provided scrape configs
+		expectedJob string
+		wantErr     bool
+	}{
+		{
+			name: "valid with spaces removed",
+			configs: []Config{
+				{
+					JobName: "jobby job job",
+				},
+			},
+			wantErr:     false,
+			expectedJob: "jobby_job_job",
+		},
+		{
+			name: "missing job",
+			configs: []Config{
+				{
+				},
+			},
+			wantErr: true,
+		},
+		{
+			name: "duplicate job",
+			configs: []Config{
+				{
+					JobName: "job1",
+				},
+				{
+					JobName: "job1",
+				},
+			},
+			wantErr: true,
+		},
+		{
+			name: "validate with special characters",
+			configs: []Config{
+				{
+					JobName: "job$1-2!3@4*job",
+				},
+			},
+			wantErr:     false,
+			expectedJob: "job_1_2_3_4_job",
+		},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			err := ValidateJobName(tt.configs)
+			if (err != nil) != tt.wantErr {
+				t.Errorf("validateJobName() error = %v, wantErr %v", err, tt.wantErr)
+				return
+			}
+			if !tt.wantErr {
+				if tt.configs[0].JobName != tt.expectedJob {
+					t.Errorf("Expected to find a job with name %v but did not find it", tt.expectedJob)
+					return
+				}
+			}
+		})
+	}
+}