Configurable Timeout

[mheripsos]

The composer-dev start ... command times out often when there are very big dependencies added it tries to install. It seems like this this timeout is not configurable, and I am not sure if it will actually affect the cloud version. See the below, which I think is just a consequence of adding a torch dependency. You can see some of these are pretty monster packages, and maybe my internet isn't great right now, but it can't finish the download/installs before the timeout.

Thanksfully, the downloads are cached, so I have to iteratively keep trying to start (actually have to re-stop and re-start each time) until it gets through all the dependencies.

Is it possible somehow to dis-activate, even temporarily, the timeout setting?

2025-08-16T01:22:46.329955634Z Using cached torch-2.8.0-cp311-cp311-manylinux_2_28_x86_64.whl (888.1 MB)
2025-08-16T01:23:06.505548387Z Using cached nvidia_cublas_cu12-12.8.4.1-py3-none-manylinux_2_27_x86_64.whl (594.3 MB)
2025-08-16T01:23:11.876885229Z Using cached nvidia_cuda_cupti_cu12-12.8.90-py3-none-manylinux2014_x86_64.manylinux_2_17_x86_64.whl (10.2 MB)
2025-08-16T01:23:13.397254486Z Using cached nvidia_cuda_nvrtc_cu12-12.8.93-py3-none-manylinux2010_x86_64.manylinux_2_12_x86_64.whl (88.0 MB)
2025-08-16T01:23:14.221635832Z Using cached nvidia_cuda_runtime_cu12-12.8.90-py3-none-manylinux2014_x86_64.manylinux_2_17_x86_64.whl (954 kB)
2025-08-16T01:23:31.289718550Z Using cached nvidia_cudnn_cu12-9.10.2.21-py3-none-manylinux_2_27_x86_64.whl (706.8 MB)
2025-08-16T01:23:40.059845640Z Using cached nvidia_cufft_cu12-11.3.3.83-py3-none-manylinux2014_x86_64.manylinux_2_17_x86_64.whl (193.1 MB)
2025-08-16T01:23:41.717145409Z Using cached nvidia_cufile_cu12-1.13.1.3-py3-none-manylinux2014_x86_64.manylinux_2_17_x86_64.whl (1.2 MB)
2025-08-16T01:23:43.187827705Z Using cached nvidia_curand_cu12-10.3.9.90-py3-none-manylinux_2_27_x86_64.whl (63.6 MB)
2025-08-16T01:23:48.788178429Z Using cached nvidia_cusolver_cu12-11.7.3.90-py3-none-manylinux_2_27_x86_64.whl (267.5 MB)
2025-08-16T01:23:56.963533911Z Using cached nvidia_cusparse_cu12-12.5.8.93-py3-none-manylinux2014_x86_64.manylinux_2_17_x86_64.whl (288.2 MB)
2025-08-16T01:24:05.097240581Z Using cached nvidia_cusparselt_cu12-0.7.1-py3-none-manylinux2014_x86_64.whl (287.2 MB)
2025-08-16T01:24:14.375125262Z Using cached nvidia_nccl_cu12-2.27.3-py3-none-manylinux2014_x86_64.manylinux_2_17_x86_64.whl (322.4 MB)
2025-08-16T01:24:19.441759369Z Downloading nvidia_nvjitlink_cu12-12.8.93-py3-none-manylinux2010_x86_64.manylinux_2_12_x86_64.whl (39.3 MB)
2025-08-16T01:25:13.157541820Z ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 39.3/39.3 MB 833.9 kB/s eta 0:00:00
2025-08-16T01:25:13.216146009Z Downloading nvidia_nvtx_cu12-12.8.90-py3-none-manylinux2014_x86_64.manylinux_2_17_x86_64.whl (89 kB)
2025-08-16T01:25:13.286217080Z ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 90.0/90.0 kB 1.4 MB/s eta 0:00:00
2025-08-16T01:25:13.450555617Z Downloading triton-3.4.0-cp311-cp311-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl (155.5 MB)
2025-08-16T01:28:27.019898369Z ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 155.5/155.5 MB 429.4 kB/s eta 0:00:00
╭─ Error ────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────╮
│ Environment did not start in 300 seconds. │
│ │
│ To print debug messages please add --debug flag. │
╰────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────╯

[ahidalgob]

Hi @mheripsos, right now we don't have a way to configure this timeout, but if you can contribute the change we're happy to accept it.

You could add a new flag to start() in cli.py like option_port, then pass it through env.start() to timeout_occurred() and there use the default one or the user-defined one if it exists.