diff --git a/router/src/infer.rs b/router/src/infer.rs
index 22131945..4973dfad 100644
--- a/router/src/infer.rs
+++ b/router/src/infer.rs
@@ -427,6 +427,7 @@ async fn decode(
             generation_health.store(true, Ordering::Relaxed);
             // Send generated tokens and filter stopped entries
             filter_send_generations(generations, entries);
+            tracing::debug!("filter batch");
 
             // Filter next batch and remove requests that were stopped
             let next_batch = filter_batch(client, next_batch, entries).await;
@@ -472,9 +473,11 @@ async fn filter_batch(
         // Next batch is now empty
         // Clear it from the Python shards cache
         // We unwrap here as we need to panic since we cannot recover if this method fails
+        tracing::info!("Call python clear cache");
         client.clear_cache(Some(id)).await.unwrap();
         None
     } else {
+        tracing::info!("Call python filter batch");
         // Filter Python shard cache
         // We unwrap here as we need to panic since we cannot recover if this method fails
         client.filter_batch(id, batch.request_ids).await.unwrap()
@@ -519,6 +522,7 @@ fn send_responses(
 ) -> Result<bool, Box<SendTimeoutError<Result<InferStreamResponse, InferError>>>> {
     // Return directly if the channel is disconnected
     if entry.response_tx.is_disconnected() {
+        tracing::debug!("Disconnected");
         metrics::increment_counter!("tgi_request_failure", "err" => "dropped");
         return Ok(true);
     }
@@ -527,6 +531,7 @@ fn send_responses(
 
     if let Some(prefill_tokens) = generation.prefill_tokens {
         // Send message
+        tracing::debug!("Send prefill");
         entry.response_tx.send_timeout(
             Ok(InferStreamResponse::Prefill(prefill_tokens)),
             Duration::from_millis(10),
@@ -543,6 +548,7 @@ fn send_responses(
 
     // generation.top_tokens
 
+    tracing::debug!("Top tokens");
     let mut top_tokens = Vec::new();
     if let Some(top_tokens_) = generation.top_tokens {
         top_tokens.extend(
@@ -565,6 +571,7 @@ fn send_responses(
         // Generation has ended
         stopped = true;
         // Send message
+        tracing::debug!("send final");
         entry.response_tx.send_timeout(
             Ok(InferStreamResponse::End {
                 token,
@@ -576,6 +583,7 @@ fn send_responses(
             Duration::from_millis(10),
         )?;
     } else {
+        tracing::debug!("send intermediate");
         // Send message
         entry.response_tx.send_timeout(
             Ok(InferStreamResponse::Intermediate { token, top_tokens }),