feat(devnet): Docker compose Otlp setup (#2216)

compose.devnet.run.yml

@@ -1,5 +1,8 @@
 version: '3.8'
 
+include:
+  - testnet/compose.tracing.yml
+
 services:
 
   explorer:
@@ -18,6 +21,7 @@ services:
     container_name: cfgsync
     image: ghcr.io/logos-blockchain/logos-blockchain:devnet
     volumes:
+      - node-data:/node-data
       - ./testnet:/etc/logos-blockchain
     depends_on:
       - grafana
@@ -109,69 +113,6 @@ services:
       - "18083:18083/tcp"
     entrypoint: /etc/logos-blockchain/scripts/run_node.sh
 
-  prometheus:
-    container_name: prometheus
-    image: prom/prometheus:v3.0.1
-    volumes:
-      - ./testnet/monitoring/prometheus.yml:/etc/prometheus/prometheus.yml:z
-    command:
-      - --config.file=/etc/prometheus/prometheus.yml
-      - --storage.tsdb.retention.time=7d
-      - --web.enable-otlp-receiver
-    ports:
-      - 0.0.0.0:9090:9090
-    restart: on-failure
-
-  grafana:
-    container_name: grafana
-    image: grafana/grafana:latest
-    environment:
-      - GF_AUTH_ANONYMOUS_ENABLED=true
-      - GF_AUTH_ANONYMOUS_ORG_ROLE=Admin
-      - GF_AUTH_DISABLE_LOGIN_FORM=true
-      - GF_FEATURE_TOGGLES_ENABLE=traceqlEditor traceQLStreaming metricsSummary
-    env_file:
-      - ./testnet/monitoring/grafana/plugins.env
-    volumes:
-      - ./testnet/monitoring/grafana/grafana.ini:/etc/grafana/grafana.ini:z
-      - ./testnet/monitoring/grafana/datasources.yaml:/etc/grafana/provisioning/datasources/datasources.yaml:z
-      - ./testnet/monitoring/grafana/dashboards.yml:/etc/grafana/provisioning/dashboards/dashboards.yml:z
-      - ./testnet/monitoring/grafana/dashboards:/var/lib/grafana/dashboards/:z
-    ports:
-      - 9091:3000
-    restart: on-failure
-    depends_on:
-      - prometheus
-
-  loki:
-    image: grafana/loki:2.9.2
-    ports:
-      - "3100:3100"
-    command: -config.file=/etc/loki/local-config.yaml
-
-  # Tempo runs as unpriviliged user, volumes need to be chowned before running.
-  tempo-init:
-    image: busybox:latest
-    user: root
-    entrypoint:
-      - "chown"
-      - "10001:10001"
-      - "/var/tempo"
-    volumes:
-      - ./tempo-data:/var/tempo
-
-  tempo:
-    image: grafana/tempo:latest
-    container_name: tempo
-    command: [ "-config.file=/etc/tempo.yaml" ]
-    volumes:
-      - ./testnet/monitoring/tempo.yaml:/etc/tempo.yaml:z
-      - ./tempo-data:/var/tempo
-    ports:
-      - "4317:4317" # otlp grpc
-    depends_on:
-      - tempo-init
-
   # L2 Sequencer Demo Services
   l2-nginx:
     container_name: l2_nginx

nodes/node/binary/src/config/tracing/serde/logger.rs

@@ -10,6 +10,7 @@ pub enum Layer {
     Gelf(GelfConfig),
     File(FileConfig),
     Loki(LokiConfig),
+    Otlp(OtlpConfig),
     #[default]
     Stdout,
     Stderr,
@@ -31,6 +32,10 @@ impl From<Layer> for LoggerLayer {
                 endpoint: config.endpoint,
                 host_identifier: config.host_identifier,
             }),
+            Layer::Otlp(config) => Self::Otlp(lb_tracing::logging::otlp::OtlpConfig {
+                endpoint: config.endpoint,
+                service_name: config.service_name,
+            }),
             Layer::Stdout => Self::Stdout,
             Layer::Stderr => Self::Stderr,
             Layer::None => Self::None,
@@ -73,3 +78,9 @@ pub struct LokiConfig {
     pub endpoint: Url,
     pub host_identifier: String,
 }
+
+#[derive(Clone, Debug, Serialize, Deserialize)]
+pub struct OtlpConfig {
+    pub endpoint: Url,
+    pub service_name: String,
+}

testnet/cfgsync.yaml

@@ -13,7 +13,9 @@ mode: Run
 
 # Tracing
 tracing_settings:
-  logger: Stdout
+  logger: !Otlp
+    endpoint: http://otel:4317
+    service_name: init_node
   tracing: None
   filter: None
   metrics: !Otlp

testnet/compose.tracing.yml

@@ -0,0 +1,67 @@
+services:
+
+  grafana:
+    container_name: grafana
+    image: grafana/grafana:latest
+    environment:
+      - GF_AUTH_ANONYMOUS_ENABLED=true
+      - GF_AUTH_ANONYMOUS_ORG_ROLE=Admin
+      - GF_AUTH_DISABLE_LOGIN_FORM=true
+      - GF_FEATURE_TOGGLES_ENABLE=traceqlEditor traceQLStreaming metricsSummary
+    env_file:
+      - ./testnet/tracing/grafana/plugins.env
+    volumes:
+      - ./testnet/tracing/grafana/grafana.ini:/etc/grafana/grafana.ini:z
+      - ./testnet/tracing/grafana/datasources.yaml:/etc/grafana/provisioning/datasources/datasources.yaml:z
+      - ./testnet/tracing/grafana/dashboards.yml:/etc/grafana/provisioning/dashboards/dashboards.yml:z
+      - ./testnet/tracing/grafana/dashboards:/var/lib/grafana/dashboards/:z
+    ports:
+      - 9091:3000
+    restart: on-failure
+    depends_on:
+      - prometheus
+      - loki
+
+  otel:
+    image: otel/opentelemetry-collector-contrib:latest
+    command: ["--config=/etc/otel-collector-config.yaml"]
+    volumes:
+      - ./testnet/tracing/otel.yaml:/etc/otel-collector-config.yaml
+    ports:
+      - "0.0.0.0:3100:4317"
+      - "14318:4318"
+
+  loki:
+    image: grafana/loki:latest
+    ports:
+      - "23100:3100"
+      - "24317:4317"
+      - "24318:4318"
+    volumes:
+      - ./testnet/tracing/loki.yaml:/etc/loki/local-config.yaml
+    command: -config.file=/etc/loki/local-config.yaml
+
+  prometheus:
+    container_name: prometheus
+    image: docker.io/prom/prometheus:v3.0.1
+    volumes:
+      - ./testnet/tracing/prometheus.yaml:/etc/prometheus/prometheus.yml:z
+    command:
+      - --config.file=/etc/prometheus/prometheus.yml
+      - --storage.tsdb.retention.time=7d
+      - --web.enable-otlp-receiver
+    ports:
+      - "0.0.0.0:9090:9090"
+    restart: on-failure
+
+  tempo:
+    image: grafana/tempo:latest
+    container_name: tempo
+    command: [ "-config.file=/etc/tempo.yaml" ]
+    volumes:
+      - ./testnet/tracing/tempo.yaml:/etc/tempo.yaml:z
+      - ./tempo-data:/var/tempo
+    ports:
+      - "34317:4317"
+    depends_on:
+      - tempo-init

testnet/tracing/loki.yaml

@@ -0,0 +1,32 @@
+auth_enabled: false
+
+server:
+  http_listen_port: 3100
+  grpc_listen_port: 9096
+
+common:
+  instance_addr: 127.0.0.1
+  path_prefix: /loki
+  storage:
+    filesystem:
+      chunks_directory: /loki/chunks
+      rules_directory: /loki/rules
+  replication_factor: 1
+  ring:
+    kvstore:
+      store: inmemory
+
+schema_config:
+  configs:
+    - from: 2024-01-01
+      store: tsdb
+      object_store: filesystem
+      schema: v13
+      index:
+        prefix: index_
+        period: 24h
+
+limits_config:
+  allow_structured_metadata: true
+  reject_old_samples: true
+  reject_old_samples_max_age: 168h

testnet/tracing/otel.yaml

@@ -0,0 +1,22 @@
+receivers:
+  otlp:
+    protocols:
+      grpc:
+        endpoint: "0.0.0.0:4317"
+      http:
+        endpoint: "0.0.0.0:4318"
+
+exporters:
+  otlp_http/loki:
+    endpoint: "http://loki:3100/otlp"
+    tls:
+      insecure: true
+
+  debug:
+    verbosity: detailed
+
+service:
+  pipelines:
+    logs:
+      receivers: [otlp]
+      exporters: [otlp_http/loki, debug]