chore(prover-service): setup metrics

charts/prover-service/templates/deployment.yaml

@@ -18,6 +18,11 @@ spec:
         - name: {{ .Chart.Name }}
           image: "{{ .Values.image.repository }}:{{ .Values.image.tag }}"
           imagePullPolicy: {{ .Values.image.pullPolicy }}
+          ports:
+            - containerPort: 9090
+              name: grpc
+            - containerPort: 4014
+              name: http-prometheus
           envFrom:
             - secretRef:
                 name: {{ if .Values.useExistingSecrets }}{{ .Values.existingSecretName }}{{ else }}{{ include "service.fullname" . }}{{ end }}

charts/prover-service/templates/service.yaml

@@ -0,0 +1,17 @@
+apiVersion: v1
+kind: Service
+metadata:
+  name: {{include "service.fullname" .}}
+  labels: {{- include "service.labels" . | nindent 4}}
+spec:
+  type: ClusterIP
+  ports:
+    - port: 9090
+      targetPort: grpc
+      protocol: TCP
+      name: grpc
+    - port: 4014
+      targetPort: http-prometheus
+      protocol: TCP
+      name: http-prometheus
+  selector: {{- include "service.selectorLabels" . | nindent 4}}

prover-service/Cargo.toml

@@ -19,6 +19,9 @@ clap = { version = "4.5", features = ["derive"] }
 hex = "0.4.3"
 log = "0.4.22"
 env_logger = "0.11"
+prometheus = "0.13"
+lazy_static = "1.4"
+warp = "0.3"
 
 [build-dependencies]
 tonic-build = "0.10"

prover-service/Dockerfile

@@ -21,6 +21,6 @@ RUN apt-get update && apt-get install -y libssl3 ca-certificates
 COPY --from=builder /usr/src/app/target/release/server /usr/local/bin/server
 COPY --from=builder /usr/src/app/elf/riscv32im-succinct-zkvm-elf /usr/local/bin/elf/
 
-EXPOSE 9090
+EXPOSE 9090 4014
 
 CMD ["server", "--elf-path", "/usr/local/bin/elf/riscv32im-succinct-zkvm-elf"]

prover-service/src/bin/server.rs

@@ -1,8 +1,10 @@
 use clap::Parser;
 use prover_service::proto::prover::prover_service_server::ProverServiceServer;
 use prover_service::service::ProverServiceImpl;
+use std::net::SocketAddr;
 use tonic::transport::Server;
 use tonic_health::server::health_reporter;
+use warp::Filter;
 
 #[derive(Parser)]
 #[command(
@@ -32,6 +34,10 @@ struct Args {
     /// Timeout for waiting for the proof in seconds
     #[arg(long, default_value = "300", env = "TIMEOUT_SECS")]
     timeout_secs: u64,
+
+    /// Port number for Prometheus metrics
+    #[arg(long, default_value = "4014", env = "METRICS_PORT")]
+    metrics_port: u16,
 }
 
 #[tokio::main]
@@ -53,6 +59,23 @@ async fn main() -> Result<(), Box<dyn std::error::Error>> {
         .set_serving::<ProverServiceServer<ProverServiceImpl>>()
         .await;
 
+    // Start Prometheus metrics endpoint
+    let metrics_addr = format!("0.0.0.0:{}", args.metrics_port);
+    let metrics_route = warp::path!("metrics").and_then(|| async {
+        use prometheus::Encoder;
+        let encoder = prometheus::TextEncoder::new();
+        let mut buffer = Vec::new();
+        encoder.encode(&prometheus::gather(), &mut buffer).unwrap();
+        Ok::<_, warp::Rejection>(String::from_utf8(buffer).unwrap())
+    });
+
+    tokio::spawn(async move {
+        log::info!("Prometheus metrics server listening on {}", metrics_addr);
+        warp::serve(metrics_route)
+            .run(metrics_addr.parse::<SocketAddr>().unwrap())
+            .await;
+    });
+
     // Serve requests
     log::info!("Health Server and GRPC Prover Server listening on {}", addr);
     Server::builder()