Panic - Qwen3 4 model (Base Image version: cuda-1.9)

[AnujShahSifthub]

System Info

zation_time="2.420641ms" queue_time="242.890392ms" inference_time="432.736853ms"}: text_embeddings_router::http::server: router/src/http/server.rs:737: Success
2026-04-04T06:38:05.263153Z INFO embed{total_time="147.840694ms" tokenization_time="803.608µs" queue_time="36.519428ms" inference_time="73.421172ms"}: text_embeddings_router::http::server: router/src/http/server.rs:737: Success
2026-04-04T06:38:06.592171Z INFO embed{total_time="963.676433ms" tokenization_time="2.14188ms" queue_time="286.734413ms" inference_time="388.875534ms"}: text_embeddings_router::http::server: router/src/http/server.rs:737: Success
2026-04-04T06:38:07.635027Z INFO embed{total_time="1.029534361s" tokenization_time="2.31388ms" queue_time="342.30999ms" inference_time="350.090667ms"}: text_embeddings_router::http::server: router/src/http/server.rs:737: Success

2026-04-04T06:38:08.661852Z INFO embed{total_time="1.0129504s" tokenization_time="2.155938ms" queue_time="332.944104ms" inference_time="365.34161ms"}: text_embeddings_router::http::server: router/src/http/server.rs:737: Success
2026-04-04T06:38:08.810618Z INFO embed{total_time="135.694009ms" tokenization_time="648.989µs" queue_time="45.665297ms" inference_time="49.182805ms"}: text_embeddings_router::http::server: router/src/http/server.rs:737: Success

thread 'tokio-runtime-worker' (61) panicked at /usr/src/backends/src/lib.rs:346:14:
No backend receiver. This is a bug.: "Full(..)"
note: run with RUST_BACKTRACE=1 environment variable to display a backtrace

Information

• Docker
• The CLI directly

Tasks

• An officially supported command
• My own modifications

Reproduction

Inside EKS Cluster on AWS on g5.2xlarge

Expected behavior

return rate limit or something instead of this and crashing