Clean up config handling in target allocator (#2470)

cmd/otel-allocator/config/config.go

@@ -35,9 +35,13 @@ import (
 	"sigs.k8s.io/controller-runtime/pkg/log/zap"
 )
 
-const DefaultResyncTime = 5 * time.Minute
-const DefaultConfigFilePath string = "/conf/targetallocator.yaml"
-const DefaultCRScrapeInterval model.Duration = model.Duration(time.Second * 30)
+const (
+	DefaultResyncTime                        = 5 * time.Minute
+	DefaultConfigFilePath     string         = "/conf/targetallocator.yaml"
+	DefaultCRScrapeInterval   model.Duration = model.Duration(time.Second * 30)
+	DefaultAllocationStrategy                = "consistent-hashing"
+	DefaultFilterStrategy                    = "relabel-config"
+)
 
 type Config struct {
 	ListenAddr             string                `yaml:"listen_addr,omitempty"`
@@ -46,8 +50,8 @@ type Config struct {
 	RootLogger             logr.Logger           `yaml:"-"`
 	CollectorSelector      *metav1.LabelSelector `yaml:"collector_selector,omitempty"`
 	PromConfig             *promconfig.Config    `yaml:"config"`
-	AllocationStrategy     *string               `yaml:"allocation_strategy,omitempty"`
-	FilterStrategy         *string               `yaml:"filter_strategy,omitempty"`
+	AllocationStrategy     string                `yaml:"allocation_strategy,omitempty"`
+	FilterStrategy         string                `yaml:"filter_strategy,omitempty"`
 	PrometheusCR           PrometheusCRConfig    `yaml:"prometheus_cr,omitempty"`
 	PodMonitorSelector     map[string]string     `yaml:"pod_monitor_selector,omitempty"`
 	ServiceMonitorSelector map[string]string     `yaml:"service_monitor_selector,omitempty"`
@@ -58,20 +62,6 @@ type PrometheusCRConfig struct {
 	ScrapeInterval model.Duration `yaml:"scrape_interval,omitempty"`
 }
 
-func (c Config) GetAllocationStrategy() string {
-	if c.AllocationStrategy != nil {
-		return *c.AllocationStrategy
-	}
-	return "consistent-hashing"
-}
-
-func (c Config) GetTargetsFilterStrategy() string {
-	if c.FilterStrategy != nil {
-		return *c.FilterStrategy
-	}
-	return "relabel-config"
-}
-
 func LoadFromFile(file string, target *Config) error {
 	return unmarshal(target, file)
 }
@@ -127,6 +117,8 @@ func unmarshal(cfg *Config, configFile string) error {
 
 func CreateDefaultConfig() Config {
 	return Config{
+		AllocationStrategy: DefaultAllocationStrategy,
+		FilterStrategy:     DefaultFilterStrategy,
 		PrometheusCR: PrometheusCRConfig{
 			ScrapeInterval: DefaultCRScrapeInterval,
 		},

cmd/otel-allocator/config/config_test.go

@@ -45,12 +45,14 @@ func TestLoad(t *testing.T) {
 				file: "./testdata/config_test.yaml",
 			},
 			want: Config{
+				AllocationStrategy: DefaultAllocationStrategy,
 				CollectorSelector: &metav1.LabelSelector{
 					MatchLabels: map[string]string{
 						"app.kubernetes.io/instance":   "default.test",
 						"app.kubernetes.io/managed-by": "opentelemetry-operator",
 					},
 				},
+				FilterStrategy: DefaultFilterStrategy,
 				PrometheusCR: PrometheusCRConfig{
 					ScrapeInterval: model.Duration(time.Second * 60),
 				},
@@ -111,12 +113,14 @@ func TestLoad(t *testing.T) {
 				file: "./testdata/pod_service_selector_test.yaml",
 			},
 			want: Config{
+				AllocationStrategy: DefaultAllocationStrategy,
 				CollectorSelector: &metav1.LabelSelector{
 					MatchLabels: map[string]string{
 						"app.kubernetes.io/instance":   "default.test",
 						"app.kubernetes.io/managed-by": "opentelemetry-operator",
 					},
 				},
+				FilterStrategy: DefaultFilterStrategy,
 				PrometheusCR: PrometheusCRConfig{
 					ScrapeInterval: DefaultCRScrapeInterval,
 				},

cmd/otel-allocator/main.go

@@ -80,8 +80,8 @@ func main() {
 	ctx := context.Background()
 	log := ctrl.Log.WithName("allocator")
 
-	allocatorPrehook = prehook.New(cfg.GetTargetsFilterStrategy(), log)
-	allocator, err = allocation.New(cfg.GetAllocationStrategy(), log, allocation.WithFilter(allocatorPrehook))
+	allocatorPrehook = prehook.New(cfg.FilterStrategy, log)
+	allocator, err = allocation.New(cfg.AllocationStrategy, log, allocation.WithFilter(allocatorPrehook))
 	if err != nil {
 		setupLog.Error(err, "Unable to initialize allocation strategy")
 		os.Exit(1)
@@ -134,11 +134,16 @@ func main() {
 	runGroup.Add(
 		func() error {
 			// Initial loading of the config file's scrape config
-			err = targetDiscoverer.ApplyConfig(allocatorWatcher.EventSourceConfigMap, cfg.PromConfig)
-			if err != nil {
-				setupLog.Error(err, "Unable to apply initial configuration")
-				return err
+			if cfg.PromConfig != nil {
+				err = targetDiscoverer.ApplyConfig(allocatorWatcher.EventSourceConfigMap, cfg.PromConfig.ScrapeConfigs)
+				if err != nil {
+					setupLog.Error(err, "Unable to apply initial configuration")
+					return err
+				}
+			} else {
+				setupLog.Info("Prometheus config empty, skipping initial discovery configuration")
 			}
+
 			err := targetDiscoverer.Watch(allocator.SetTargets)
 			setupLog.Info("Target discoverer exited")
 			return err
@@ -180,7 +185,7 @@ func main() {
 						setupLog.Error(err, "Unable to load configuration")
 						continue
 					}
-					err = targetDiscoverer.ApplyConfig(event.Source, loadConfig)
+					err = targetDiscoverer.ApplyConfig(event.Source, loadConfig.ScrapeConfigs)
 					if err != nil {
 						setupLog.Error(err, "Unable to apply configuration")
 						continue

cmd/otel-allocator/target/discovery.go

@@ -22,7 +22,7 @@ import (
 	"github.com/prometheus/client_golang/prometheus"
 	"github.com/prometheus/client_golang/prometheus/promauto"
 	"github.com/prometheus/common/model"
-	"github.com/prometheus/prometheus/config"
+	promconfig "github.com/prometheus/prometheus/config"
 	"github.com/prometheus/prometheus/discovery"
 	"github.com/prometheus/prometheus/model/relabel"
 	"gopkg.in/yaml.v3"
@@ -41,7 +41,7 @@ type Discoverer struct {
 	log                  logr.Logger
 	manager              *discovery.Manager
 	close                chan struct{}
-	configsMap           map[allocatorWatcher.EventSource]*config.Config
+	configsMap           map[allocatorWatcher.EventSource][]*promconfig.ScrapeConfig
 	hook                 discoveryHook
 	scrapeConfigsHash    hash.Hash
 	scrapeConfigsUpdater scrapeConfigsUpdater
@@ -52,33 +52,30 @@ type discoveryHook interface {
 }
 
 type scrapeConfigsUpdater interface {
-	UpdateScrapeConfigResponse(map[string]*config.ScrapeConfig) error
+	UpdateScrapeConfigResponse(map[string]*promconfig.ScrapeConfig) error
 }
 
 func NewDiscoverer(log logr.Logger, manager *discovery.Manager, hook discoveryHook, scrapeConfigsUpdater scrapeConfigsUpdater) *Discoverer {
 	return &Discoverer{
 		log:                  log,
 		manager:              manager,
 		close:                make(chan struct{}),
-		configsMap:           make(map[allocatorWatcher.EventSource]*config.Config),
+		configsMap:           make(map[allocatorWatcher.EventSource][]*promconfig.ScrapeConfig),
 		hook:                 hook,
+		scrapeConfigsHash:    nil, // we want the first update to succeed even if the config is empty
 		scrapeConfigsUpdater: scrapeConfigsUpdater,
 	}
 }
 
-func (m *Discoverer) ApplyConfig(source allocatorWatcher.EventSource, cfg *config.Config) error {
-	if cfg == nil {
-		m.log.Info("Service Discovery got empty Prometheus config", "source", source.String())
-		return nil
-	}
-	m.configsMap[source] = cfg
-	jobToScrapeConfig := make(map[string]*config.ScrapeConfig)
+func (m *Discoverer) ApplyConfig(source allocatorWatcher.EventSource, scrapeConfigs []*promconfig.ScrapeConfig) error {
+	m.configsMap[source] = scrapeConfigs
+	jobToScrapeConfig := make(map[string]*promconfig.ScrapeConfig)
 
 	discoveryCfg := make(map[string]discovery.Configs)
 	relabelCfg := make(map[string][]*relabel.Config)
 
-	for _, value := range m.configsMap {
-		for _, scrapeConfig := range value.ScrapeConfigs {
+	for _, configs := range m.configsMap {
+		for _, scrapeConfig := range configs {
 			jobToScrapeConfig[scrapeConfig.JobName] = scrapeConfig
 			discoveryCfg[scrapeConfig.JobName] = scrapeConfig.ServiceDiscoveryConfigs
 			relabelCfg[scrapeConfig.JobName] = scrapeConfig.RelabelConfigs
@@ -137,7 +134,7 @@ func (m *Discoverer) Close() {
 
 // Calculate a hash for a scrape config map.
 // This is done by marshaling to YAML because it's the most straightforward and doesn't run into problems with unexported fields.
-func getScrapeConfigHash(jobToScrapeConfig map[string]*config.ScrapeConfig) (hash.Hash64, error) {
+func getScrapeConfigHash(jobToScrapeConfig map[string]*promconfig.ScrapeConfig) (hash.Hash64, error) {
 	var err error
 	hash := fnv.New64()
 	yamlEncoder := yaml.NewEncoder(hash)

cmd/otel-allocator/target/discovery_test.go

@@ -89,7 +89,7 @@ func TestDiscovery(t *testing.T) {
 			err := config.LoadFromFile(tt.args.file, &cfg)
 			assert.NoError(t, err)
 			assert.True(t, len(cfg.PromConfig.ScrapeConfigs) > 0)
-			err = manager.ApplyConfig(allocatorWatcher.EventSourcePrometheusCR, cfg.PromConfig)
+			err = manager.ApplyConfig(allocatorWatcher.EventSourcePrometheusCR, cfg.PromConfig.ScrapeConfigs)
 			assert.NoError(t, err)
 
 			gotTargets := <-results
@@ -306,11 +306,11 @@ func TestDiscovery_ScrapeConfigHashing(t *testing.T) {
 
 	for _, tc := range tests {
 		t.Run(tc.description, func(t *testing.T) {
-			err := manager.ApplyConfig(allocatorWatcher.EventSourcePrometheusCR, tc.cfg)
+			err := manager.ApplyConfig(allocatorWatcher.EventSourcePrometheusCR, tc.cfg.ScrapeConfigs)
 			if !tc.expectErr {
 				expectedConfig = make(map[string]*promconfig.ScrapeConfig)
-				for _, value := range manager.configsMap {
-					for _, scrapeConfig := range value.ScrapeConfigs {
+				for _, configs := range manager.configsMap {
+					for _, scrapeConfig := range configs {
 						expectedConfig[scrapeConfig.JobName] = scrapeConfig
 					}
 				}
@@ -389,7 +389,7 @@ func BenchmarkApplyScrapeConfig(b *testing.B) {
 
 	b.ResetTimer()
 	for i := 0; i < b.N; i++ {
-		err := manager.ApplyConfig(allocatorWatcher.EventSourcePrometheusCR, cfg)
+		err := manager.ApplyConfig(allocatorWatcher.EventSourcePrometheusCR, cfg.ScrapeConfigs)
 		require.NoError(b, err)
 	}
 }