diff --git a/Dockerfile b/Dockerfile
index 592f1f72..a94bc3a0 100644
--- a/Dockerfile
+++ b/Dockerfile
@@ -37,7 +37,7 @@ ENV LANG=C.UTF-8 \
     MODEL_ID=bigscience/bloom-560m \
     QUANTIZE=false \
     NUM_SHARD=1 \
-    PORT=80 \
+    PORT=8080 \
     CUDA_HOME=/usr/local/cuda \
     LD_LIBRARY_PATH="/opt/miniconda/envs/text-generation/lib:/usr/local/cuda/lib64:/usr/local/cuda/extras/CUPTI/lib64:$LD_LIBRARY_PATH" \
     CONDA_DEFAULT_ENV=text-generation \
diff --git a/launcher/src/main.rs b/launcher/src/main.rs
index 80466fe6..f015a393 100644
--- a/launcher/src/main.rs
+++ b/launcher/src/main.rs
@@ -19,13 +19,13 @@ use subprocess::{ExitStatus, Popen, PopenConfig, PopenError, Redirection};
 #[derive(Parser, Debug)]
 #[clap(author, version, about, long_about = None)]
 struct Args {
-    #[clap(default_value = "bigscience/bloom-560m", long, env)]
+    #[clap(default_value = "bigscience/bloom-560m", long, env = "HF_MODEL_ID")]
     model_id: String,
     #[clap(long, env)]
     revision: Option<String>,
     #[clap(long, env)]
     sharded: Option<bool>,
-    #[clap(long, env)]
+    #[clap(long, env = "SM_NUM_GPUS")]
     num_shard: Option<usize>,
     #[clap(long, env)]
     quantize: bool,
diff --git a/router/src/server.rs b/router/src/server.rs
index 3b63ec8a..5e0535c6 100644
--- a/router/src/server.rs
+++ b/router/src/server.rs
@@ -530,10 +530,12 @@ pub async fn run(
     let app = Router::new()
         .merge(SwaggerUi::new("/docs").url("/api-doc/openapi.json", ApiDoc::openapi()))
         .route("/", post(compat_generate))
+        .route("/invocations", post(compat_generate))
         .route("/generate", post(generate))
         .route("/generate_stream", post(generate_stream))
         .route("/", get(health))
         .route("/health", get(health))
+        .route("/ping", get(health))
         .route("/metrics", get(metrics))
         .layer(Extension(compat_return_full_text))
         .layer(Extension(infer))