Merge branch 'main' into kent.2.7.last_changes

app/app.go

@@ -25,28 +25,31 @@ type App struct {
 // Start exits, Stop will be called on all dependencies then on App then the
 // program will exit.
 func (a *App) Start() error {
+	// little helper function to record the current config and rules hashes; we call it in
+	// the callback but also at startup
+	record_hashes := func(msg string) {
+		cfgHash, rulesHash := a.Config.GetHashes()
+		if a.Logger != nil {
+			a.Logger.Warn().WithFields(map[string]interface{}{
+				"configHash": cfgHash,
+				"rulesHash":  rulesHash,
+			}).Logf(msg)
+		}
+		cfgMetric := config.ConfigHashMetrics(cfgHash)
+		ruleMetric := config.ConfigHashMetrics(rulesHash)
+		a.Metrics.Gauge("config_hash", cfgMetric)
+		a.Metrics.Gauge("rule_config_hash", ruleMetric)
+	}
+
 	a.Logger.Debug().Logf("Starting up App...")
 	a.Metrics.Register("config_hash", "gauge")
 	a.Metrics.Register("rule_config_hash", "gauge")
-
 	a.IncomingRouter.SetVersion(a.Version)
 	a.PeerRouter.SetVersion(a.Version)
 
+	record_hashes("loaded configuration at startup")
 	a.Config.RegisterReloadCallback(func(configHash, rulesHash string) {
-		if a.Logger != nil {
-			a.Logger.Warn().WithFields(map[string]interface{}{
-				"configHash": configHash,
-				"rulesHash":  rulesHash,
-			}).Logf("configuration change was detected and the configuration was reloaded.")
-
-			cfgMetric := config.ConfigHashMetrics(configHash)
-			ruleMetric := config.ConfigHashMetrics(rulesHash)
-
-			a.Metrics.Gauge("config_hash", cfgMetric)
-			a.Metrics.Gauge("rule_config_hash", ruleMetric)
-
-		}
-
+		record_hashes("configuration change was detected and the configuration was reloaded.")
 	})
 
 	// launch our main routers to listen for incoming event traffic from both peers

cmd/refinery/main.go

@@ -116,6 +116,7 @@ func main() {
 		os.Exit(1)
 	}
 
+	// set up the peer management and pubsub implementations
 	var peers peer.Peers
 	var pubsubber pubsub.PubSub
 	ptype, err := c.GetPeerManagementType()
@@ -124,24 +125,15 @@ func main() {
 	}
 	switch ptype {
 	case "file":
-		// we want to use the file peers object to ask if we have only one peer, so we need to instantiate it
-		// with a dummy metrics object. we'll replace it with the real one later.
-		peers = &peer.FilePeers{Cfg: c, Metrics: &metrics.NullMetrics{}}
-		// if we only have one, we can use the local pubsub implementation.
-		peerList, err := peers.GetPeers()
-		if err != nil {
-			panic(err)
-		}
-		if len(peerList) == 1 {
-			pubsubber = &pubsub.LocalPubSub{}
-		} else {
-			pubsubber = &pubsub.GoRedisPubSub{}
-		}
-		// now erase the peers Metrics object so that it will get injected with the right one later
-		peers.(*peer.FilePeers).Metrics = nil
+		// In the case of file peers, we do not use Redis for anything, including pubsub, so
+		// we use the local pubsub implementation. Even if we have multiple peers, these
+		// peers cannot communicate using pubsub.
+		peers = &peer.FilePeers{}
+		pubsubber = &pubsub.LocalPubSub{}
 	case "redis":
-		pubsubber = &pubsub.GoRedisPubSub{}
+		// if we're using redis, we need to set it up for both peers and pubsub
 		peers = &peer.RedisPubsubPeers{}
+		pubsubber = &pubsub.GoRedisPubSub{}
 	default:
 		// this should have been caught by validation
 		panic("invalid config option 'PeerManagement.Type'")

config/metadata/configMeta.yaml

@@ -835,12 +835,15 @@ groups:
           Peer management is the mechanism by which Refinery locates its peers.
 
           `file` means that Refinery gets its peer list from the Peers list in
-          this config file.
+          this config file. It also prevents Refinery from using a publish/subscribe
+          mechanism to propagate peer lists, stress levels, and configuration changes.
 
           `redis` means that Refinery uses a Publish/Subscribe mechanism,
-          implemented on Redis, to propagate peer lists much more quickly than
-          the legacy mechanism. This is the recommended setting, especially for
-          new installations.
+          implemented on Redis, to propagate peer lists, stress levels, and
+          notification of configuration changes much more quickly than the
+          legacy mechanism. This is the recommended setting, especially for new
+          installations. If this is specified, fields in `RedisPeerManagement`
+          must also be set.
 
       - name: Identifier
         v1group: PeerManagement

test/config.yaml

@@ -8,18 +8,18 @@ Network:
 
 Logger:
   Type: stdout
-  Level: debug
+  Level: info
 
 LegacyMetrics:
   Enabled: true
   Dataset: refinery_metrics
-  APIKey: HTab42OXw838zcJDH3eJEH
+  APIKey: $REFINERY_HONEYCOMB_API_KEY
   APIHost: https://api-dogfood.honeycomb.io
 
 OTelMetrics:
   Enabled: true
   Dataset: refinery_metrics_otel
-  APIKey: HTab42OXw838zcJDH3eJEH
+  APIKey: $REFINERY_HONEYCOMB_API_KEY
   APIHost: https://api-dogfood.honeycomb.io
 
 RefineryTelemetry: