diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml
index 88c90df..b59e7b0 100644
--- a/.github/workflows/ci.yml
+++ b/.github/workflows/ci.yml
@@ -80,7 +80,7 @@ jobs:
           path: |
             ~/.cargo/registry
             ~/.cargo/git
-
+      - run: bash ./ci/prepare_build_environment.sh
       - name: Bulid release binary
         run: cargo build --release --target ${{ matrix.target }}
 
diff --git a/Cargo.lock b/Cargo.lock
index 149a2d1..19dea27 100644
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -142,6 +142,28 @@ version = "1.6.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "bddcadddf5e9015d310179a59bb28c4d4b9920ad0f11e8e14dbadf654890c9a6"
 
+[[package]]
+name = "async-stream"
+version = "0.3.5"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "cd56dd203fef61ac097dd65721a419ddccb106b2d2b70ba60a6b529f03961a51"
+dependencies = [
+ "async-stream-impl",
+ "futures-core",
+ "pin-project-lite",
+]
+
+[[package]]
+name = "async-stream-impl"
+version = "0.3.5"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "16e62a023e7c117e27523144c5d2459f4397fcc3cab0085af8e2224f643a0193"
+dependencies = [
+ "proc-macro2",
+ "quote",
+ "syn 2.0.18",
+]
+
 [[package]]
 name = "async-trait"
 version = "0.1.68"
@@ -208,6 +230,22 @@ dependencies = [
  "tower-service",
 ]
 
+[[package]]
+name = "axum-tracing-opentelemetry"
+version = "0.10.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "164b95427e83b79583c7699a72b4a6b485a12bbdef5b5c054ee5ff2296d82f52"
+dependencies = [
+ "axum",
+ "futures",
+ "http",
+ "opentelemetry",
+ "tower",
+ "tower-http 0.3.5",
+ "tracing",
+ "tracing-opentelemetry",
+]
+
 [[package]]
 name = "backtrace"
 version = "0.3.67"
@@ -868,6 +906,12 @@ dependencies = [
  "windows-sys 0.48.0",
 ]
 
+[[package]]
+name = "fixedbitset"
+version = "0.4.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "0ce7134b9999ecaf8bcd65542e436736ef32ddca1b3e06094cb6ec5755203b80"
+
 [[package]]
 name = "flate2"
 version = "1.0.26"
@@ -918,6 +962,21 @@ dependencies = [
  "winapi",
 ]
 
+[[package]]
+name = "futures"
+version = "0.3.28"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "23342abe12aba583913b2e62f22225ff9c950774065e4bfb61a19cd9770fec40"
+dependencies = [
+ "futures-channel",
+ "futures-core",
+ "futures-executor",
+ "futures-io",
+ "futures-sink",
+ "futures-task",
+ "futures-util",
+]
+
 [[package]]
 name = "futures-channel"
 version = "0.3.28"
@@ -925,6 +984,7 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "955518d47e09b25bbebc7a18df10b81f0c766eaf4c4f1cccef2fca5f2a4fb5f2"
 dependencies = [
  "futures-core",
+ "futures-sink",
 ]
 
 [[package]]
@@ -933,6 +993,17 @@ version = "0.3.28"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "4bca583b7e26f571124fe5b7561d49cb2868d79116cfa0eefce955557c6fee8c"
 
+[[package]]
+name = "futures-executor"
+version = "0.3.28"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "ccecee823288125bd88b4d7f565c9e58e41858e47ab72e8ea2d64e93624386e0"
+dependencies = [
+ "futures-core",
+ "futures-task",
+ "futures-util",
+]
+
 [[package]]
 name = "futures-io"
 version = "0.3.28"
@@ -968,6 +1039,7 @@ version = "0.3.28"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "26b01e40b772d54cf6c6d721c1d1abd0647a0106a12ecaa1c186273392a69533"
 dependencies = [
+ "futures-channel",
  "futures-core",
  "futures-io",
  "futures-macro",
@@ -1153,6 +1225,18 @@ dependencies = [
  "want",
 ]
 
+[[package]]
+name = "hyper-timeout"
+version = "0.4.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "bbb958482e8c7be4bc3cf272a766a2b0bf1a6755e7a6ae777f017a31d11b13b1"
+dependencies = [
+ "hyper",
+ "pin-project-lite",
+ "tokio",
+ "tokio-io-timeout",
+]
+
 [[package]]
 name = "hyper-tls"
 version = "0.5.0"
@@ -1592,6 +1676,12 @@ dependencies = [
  "syn 2.0.18",
 ]
 
+[[package]]
+name = "multimap"
+version = "0.8.3"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "e5ce46fe64a9d73be07dcbe690a38ce1b293be448fd8ce1e6c1b8062c9f72c6a"
+
 [[package]]
 name = "murmurhash32"
 version = "0.2.0"
@@ -1769,6 +1859,86 @@ dependencies = [
  "vcpkg",
 ]
 
+[[package]]
+name = "opentelemetry"
+version = "0.18.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "69d6c3d7288a106c0a363e4b0e8d308058d56902adefb16f4936f417ffef086e"
+dependencies = [
+ "opentelemetry_api",
+ "opentelemetry_sdk",
+]
+
+[[package]]
+name = "opentelemetry-otlp"
+version = "0.11.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "d1c928609d087790fc936a1067bdc310ae702bdf3b090c3f281b713622c8bbde"
+dependencies = [
+ "async-trait",
+ "futures",
+ "futures-util",
+ "http",
+ "opentelemetry",
+ "opentelemetry-proto",
+ "prost",
+ "thiserror",
+ "tokio",
+ "tonic",
+]
+
+[[package]]
+name = "opentelemetry-proto"
+version = "0.1.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "d61a2f56df5574508dd86aaca016c917489e589ece4141df1b5e349af8d66c28"
+dependencies = [
+ "futures",
+ "futures-util",
+ "opentelemetry",
+ "prost",
+ "tonic",
+ "tonic-build",
+]
+
+[[package]]
+name = "opentelemetry_api"
+version = "0.18.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "c24f96e21e7acc813c7a8394ee94978929db2bcc46cf6b5014fc612bf7760c22"
+dependencies = [
+ "fnv",
+ "futures-channel",
+ "futures-util",
+ "indexmap",
+ "js-sys",
+ "once_cell",
+ "pin-project-lite",
+ "thiserror",
+]
+
+[[package]]
+name = "opentelemetry_sdk"
+version = "0.18.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "1ca41c4933371b61c2a2f214bf16931499af4ec90543604ec828f7a625c09113"
+dependencies = [
+ "async-trait",
+ "crossbeam-channel",
+ "dashmap",
+ "fnv",
+ "futures-channel",
+ "futures-executor",
+ "futures-util",
+ "once_cell",
+ "opentelemetry_api",
+ "percent-encoding",
+ "rand",
+ "thiserror",
+ "tokio",
+ "tokio-stream",
+]
+
 [[package]]
 name = "overload"
 version = "0.1.1"
@@ -1842,6 +2012,16 @@ version = "2.2.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "478c572c3d73181ff3c2539045f6eb99e5491218eae919370993b890cdbdd98e"
 
+[[package]]
+name = "petgraph"
+version = "0.6.3"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "4dd7d28ee937e54fe3080c91faa1c3a46c06de6252988a7f4592ba2310ef22a4"
+dependencies = [
+ "fixedbitset",
+ "indexmap",
+]
+
 [[package]]
 name = "pin-project"
 version = "1.1.0"
@@ -1901,6 +2081,16 @@ version = "0.2.17"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "5b40af805b3121feab8a3c29f04d8ad262fa8e0561883e7653e024ae4479e6de"
 
+[[package]]
+name = "prettyplease"
+version = "0.1.25"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "6c8646e95016a7a6c4adea95bafa8a16baab64b583356217f2c85db4a39d9a86"
+dependencies = [
+ "proc-macro2",
+ "syn 1.0.109",
+]
+
 [[package]]
 name = "proc-macro-error"
 version = "1.0.4"
@@ -1934,6 +2124,60 @@ dependencies = [
  "unicode-ident",
 ]
 
+[[package]]
+name = "prost"
+version = "0.11.9"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "0b82eaa1d779e9a4bc1c3217db8ffbeabaae1dca241bf70183242128d48681cd"
+dependencies = [
+ "bytes",
+ "prost-derive",
+]
+
+[[package]]
+name = "prost-build"
+version = "0.11.9"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "119533552c9a7ffacc21e099c24a0ac8bb19c2a2a3f363de84cd9b844feab270"
+dependencies = [
+ "bytes",
+ "heck",
+ "itertools 0.10.5",
+ "lazy_static",
+ "log",
+ "multimap",
+ "petgraph",
+ "prettyplease",
+ "prost",
+ "prost-types",
+ "regex",
+ "syn 1.0.109",
+ "tempfile",
+ "which",
+]
+
+[[package]]
+name = "prost-derive"
+version = "0.11.9"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "e5d2d8d10f3c6ded6da8b05b5fb3b8a5082514344d56c9f871412d29b4e075b4"
+dependencies = [
+ "anyhow",
+ "itertools 0.10.5",
+ "proc-macro2",
+ "quote",
+ "syn 1.0.109",
+]
+
+[[package]]
+name = "prost-types"
+version = "0.11.9"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "213622a1460818959ac1181aaeb2dc9c7f63df720db7d788b3e24eacd1983e13"
+dependencies = [
+ "prost",
+]
+
 [[package]]
 name = "quote"
 version = "1.0.28"
@@ -2513,11 +2757,14 @@ name = "tabby"
 version = "0.1.0"
 dependencies = [
  "axum",
+ "axum-tracing-opentelemetry",
  "clap",
  "ctranslate2-bindings",
  "hyper",
  "lazy_static",
  "mime_guess",
+ "opentelemetry",
+ "opentelemetry-otlp",
  "rust-embed",
  "serde",
  "serde_json",
@@ -2529,8 +2776,9 @@ dependencies = [
  "tabby-scheduler",
  "tokio",
  "tower",
- "tower-http",
+ "tower-http 0.4.0",
  "tracing",
+ "tracing-opentelemetry",
  "tracing-subscriber 0.3.17",
  "utoipa",
  "utoipa-swagger-ui",
@@ -2843,6 +3091,16 @@ dependencies = [
  "windows-sys 0.48.0",
 ]
 
+[[package]]
+name = "tokio-io-timeout"
+version = "1.2.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "30b74022ada614a1b4834de765f9bb43877f910cc8ce4be40e89042c9223a8bf"
+dependencies = [
+ "pin-project-lite",
+ "tokio",
+]
+
 [[package]]
 name = "tokio-macros"
 version = "2.1.0"
@@ -2864,6 +3122,17 @@ dependencies = [
  "tokio",
 ]
 
+[[package]]
+name = "tokio-stream"
+version = "0.1.14"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "397c988d37662c7dda6d2208364a706264bf3d6138b11d436cbac0ad38832842"
+dependencies = [
+ "futures-core",
+ "pin-project-lite",
+ "tokio",
+]
+
 [[package]]
 name = "tokio-util"
 version = "0.7.8"
@@ -2912,6 +3181,51 @@ dependencies = [
  "winnow",
 ]
 
+[[package]]
+name = "tonic"
+version = "0.8.3"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "8f219fad3b929bef19b1f86fbc0358d35daed8f2cac972037ac0dc10bbb8d5fb"
+dependencies = [
+ "async-stream",
+ "async-trait",
+ "axum",
+ "base64 0.13.1",
+ "bytes",
+ "futures-core",
+ "futures-util",
+ "h2",
+ "http",
+ "http-body",
+ "hyper",
+ "hyper-timeout",
+ "percent-encoding",
+ "pin-project",
+ "prost",
+ "prost-derive",
+ "tokio",
+ "tokio-stream",
+ "tokio-util",
+ "tower",
+ "tower-layer",
+ "tower-service",
+ "tracing",
+ "tracing-futures",
+]
+
+[[package]]
+name = "tonic-build"
+version = "0.8.4"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "5bf5e9b9c0f7e0a7c027dcfaba7b2c60816c7049171f679d99ee2ff65d0de8c4"
+dependencies = [
+ "prettyplease",
+ "proc-macro2",
+ "prost-build",
+ "quote",
+ "syn 1.0.109",
+]
+
 [[package]]
 name = "tower"
 version = "0.4.13"
@@ -2920,9 +3234,32 @@ checksum = "b8fa9be0de6cf49e536ce1851f987bd21a43b771b09473c3549a6c853db37c1c"
 dependencies = [
  "futures-core",
  "futures-util",
+ "indexmap",
  "pin-project",
  "pin-project-lite",
+ "rand",
+ "slab",
  "tokio",
+ "tokio-util",
+ "tower-layer",
+ "tower-service",
+ "tracing",
+]
+
+[[package]]
+name = "tower-http"
+version = "0.3.5"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "f873044bf02dd1e8239e9c1293ea39dad76dc594ec16185d0a1bf31d8dc8d858"
+dependencies = [
+ "bitflags",
+ "bytes",
+ "futures-core",
+ "futures-util",
+ "http",
+ "http-body",
+ "http-range-header",
+ "pin-project-lite",
  "tower-layer",
  "tower-service",
  "tracing",
@@ -2992,6 +3329,16 @@ dependencies = [
  "valuable",
 ]
 
+[[package]]
+name = "tracing-futures"
+version = "0.2.5"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "97d095ae15e245a057c8e8451bab9b3ee1e1f68e9ba2b4fbc18d0ac5237835f2"
+dependencies = [
+ "pin-project",
+ "tracing",
+]
+
 [[package]]
 name = "tracing-log"
 version = "0.1.3"
@@ -3003,6 +3350,20 @@ dependencies = [
  "tracing-core",
 ]
 
+[[package]]
+name = "tracing-opentelemetry"
+version = "0.18.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "21ebb87a95ea13271332df069020513ab70bdb5637ca42d6e492dc3bbbad48de"
+dependencies = [
+ "once_cell",
+ "opentelemetry",
+ "tracing",
+ "tracing-core",
+ "tracing-log",
+ "tracing-subscriber 0.3.17",
+]
+
 [[package]]
 name = "tracing-serde"
 version = "0.1.3"
@@ -3398,6 +3759,17 @@ dependencies = [
  "wasm-bindgen",
 ]
 
+[[package]]
+name = "which"
+version = "4.4.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "2441c784c52b289a054b7201fc93253e288f094e2f4be9058343127c4226a269"
+dependencies = [
+ "either",
+ "libc",
+ "once_cell",
+]
+
 [[package]]
 name = "winapi"
 version = "0.3.9"
diff --git a/Dockerfile b/Dockerfile
index c09d74a..8461764 100644
--- a/Dockerfile
+++ b/Dockerfile
@@ -9,6 +9,7 @@ RUN apt-get update && \
         curl \
         pkg-config \
         libssl-dev \
+        protobuf-compiler \
         && \
     apt-get clean && \
     rm -rf /var/lib/apt/lists/*
diff --git a/ci/prepare_build_environment.sh b/ci/prepare_build_environment.sh
new file mode 100755
index 0000000..e1b3c09
--- /dev/null
+++ b/ci/prepare_build_environment.sh
@@ -0,0 +1,5 @@
+#!/bin/bash
+
+if [[ "$OSTYPE" == "darwin"* ]]; then
+  brew install protobuf
+fi
diff --git a/crates/tabby/Cargo.toml b/crates/tabby/Cargo.toml
index 9714b15..1d53364 100644
--- a/crates/tabby/Cargo.toml
+++ b/crates/tabby/Cargo.toml
@@ -26,6 +26,11 @@ strum = { version = "0.24", features = ["derive"] }
 strfmt = "0.2.4"
 tracing = { workspace = true }
 tracing-subscriber = { workspace = true }
+opentelemetry = { version = "0.18.0", features = ["rt-tokio"] }
+opentelemetry-otlp = "0.11.0"
+axum-tracing-opentelemetry = "0.10.0"
+tracing-opentelemetry = "0.18.0"
+
 
 [dependencies.uuid]
 version = "1.3.3"
diff --git a/crates/tabby/src/main.rs b/crates/tabby/src/main.rs
index d48a9e1..4891ef9 100644
--- a/crates/tabby/src/main.rs
+++ b/crates/tabby/src/main.rs
@@ -2,7 +2,13 @@ mod download;
 mod serve;
 
 use clap::{Parser, Subcommand};
-use tracing_subscriber::EnvFilter;
+use opentelemetry::{
+    global,
+    sdk::{propagation::TraceContextPropagator, trace, trace::Sampler, Resource},
+    KeyValue,
+};
+use opentelemetry_otlp::WithExportConfig;
+use tracing_subscriber::{layer::SubscriberExt, util::SubscriberInitExt, EnvFilter, Layer};
 
 #[derive(Parser)]
 #[command(author, version, about, long_about = None)]
@@ -10,6 +16,10 @@ use tracing_subscriber::EnvFilter;
 struct Cli {
     #[command(subcommand)]
     command: Commands,
+
+    /// Open Telemetry endpoint.
+    #[clap(long)]
+    otlp_endpoint: Option<String>,
 }
 
 #[derive(Subcommand)]
@@ -33,11 +43,8 @@ pub struct SchedulerArgs {
 
 #[tokio::main]
 async fn main() {
-    tracing_subscriber::fmt::fmt()
-        .with_env_filter(EnvFilter::from_default_env().add_directive("tabby=info".parse().unwrap()))
-        .init();
-
     let cli = Cli::parse();
+    init_logging(cli.otlp_endpoint);
 
     match &cli.command {
         Commands::Serve(args) => serve::main(args).await,
@@ -46,6 +53,8 @@ async fn main() {
             .await
             .unwrap_or_else(|err| fatal!("Scheduler failed due to '{}'", err)),
     }
+
+    opentelemetry::global::shutdown_tracer_provider();
 }
 
 #[macro_export]
@@ -64,3 +73,50 @@ macro_rules! fatal {
         })
     };
 }
+
+fn init_logging(otlp_endpoint: Option<String>) {
+    let mut layers = Vec::new();
+
+    let fmt_layer = tracing_subscriber::fmt::layer()
+        .with_file(true)
+        .with_line_number(true)
+        .boxed();
+
+    layers.push(fmt_layer);
+
+    if let Some(otlp_endpoint) = &otlp_endpoint {
+        global::set_text_map_propagator(TraceContextPropagator::new());
+
+        let tracer = opentelemetry_otlp::new_pipeline()
+            .tracing()
+            .with_exporter(
+                opentelemetry_otlp::new_exporter()
+                    .tonic()
+                    .with_endpoint(otlp_endpoint),
+            )
+            .with_trace_config(
+                trace::config()
+                    .with_resource(Resource::new(vec![KeyValue::new(
+                        "service.name",
+                        "tabby.server",
+                    )]))
+                    .with_sampler(Sampler::AlwaysOn),
+            )
+            .install_batch(opentelemetry::runtime::Tokio);
+
+        if let Ok(tracer) = tracer {
+            layers.push(tracing_opentelemetry::layer().with_tracer(tracer).boxed());
+            axum_tracing_opentelemetry::init_propagator().unwrap();
+        };
+    }
+
+    let env_filter = EnvFilter::from_default_env()
+        .add_directive("tabby=info".parse().unwrap())
+        .add_directive("axum_tracing_opentelemetry=info".parse().unwrap())
+        .add_directive("otel=debug".parse().unwrap());
+
+    tracing_subscriber::registry()
+        .with(layers)
+        .with(env_filter)
+        .init();
+}
diff --git a/crates/tabby/src/serve/completions.rs b/crates/tabby/src/serve/completions.rs
index 5d67d65..21d7674 100644
--- a/crates/tabby/src/serve/completions.rs
+++ b/crates/tabby/src/serve/completions.rs
@@ -8,6 +8,7 @@ use hyper::StatusCode;
 use serde::{Deserialize, Serialize};
 use strfmt::{strfmt, strfmt_builder};
 use tabby_common::{events, path::ModelDir};
+use tracing::instrument;
 use utoipa::ToSchema;
 
 use self::languages::get_stop_words;
@@ -64,6 +65,7 @@ pub struct CompletionResponse {
         (status = 400, description = "Bad Request")
     )
 )]
+#[instrument(skip(state, request))]
 pub async fn completion(
     State(state): State<Arc<CompletionState>>,
     Json(request): Json<CompletionRequest>,
diff --git a/crates/tabby/src/serve/mod.rs b/crates/tabby/src/serve/mod.rs
index 1e622e9..df380d0 100644
--- a/crates/tabby/src/serve/mod.rs
+++ b/crates/tabby/src/serve/mod.rs
@@ -8,6 +8,7 @@ use std::{
 };
 
 use axum::{routing, Router, Server};
+use axum_tracing_opentelemetry::opentelemetry_tracing_layer;
 use clap::Args;
 use tower_http::cors::CorsLayer;
 use tracing::info;
@@ -92,8 +93,7 @@ pub async fn main(args: &ServeArgs) {
     let app = Router::new()
         .merge(SwaggerUi::new("/swagger-ui").url("/api-docs/openapi.json", ApiDoc::openapi()))
         .nest("/v1", api_router(args))
-        .fallback(fallback(args.experimental_admin_panel))
-        .layer(CorsLayer::permissive());
+        .fallback(fallback(args.experimental_admin_panel));
 
     let address = SocketAddr::from((Ipv4Addr::UNSPECIFIED, args.port));
     info!("Listening at {}", address);
@@ -112,6 +112,8 @@ fn api_router(args: &ServeArgs) -> Router {
             routing::post(completions::completion)
                 .with_state(Arc::new(completions::CompletionState::new(args))),
         )
+        .layer(CorsLayer::permissive())
+        .layer(opentelemetry_tracing_layer())
 }
 
 fn fallback(experimental_admin_panel: bool) -> routing::MethodRouter {