Not able to deploy my langgraph agent in python

[EHTASHAM_TOOR]

i am facing this error in the picture. tried to redoploy again but same issue comes up.

[David]

Hi, we are looking into it. We will get back to you on this.

Have you tried with a bigger instance than the Free instance?

[Bjarnos]

I had this issue before, I believe it’s like David mentioned.
You are trying to download a (docker) image that is bigger than your instance allows. If you own the image you could try minimizing it. Otherwise, I’d recommend either upgrading your instance or (if that’s not an option) looking into a different host method with more generous resource limits.

[EHTASHAM_TOOR]

Thanks for looking into it.
Okay got it. Can you tell me whats the available image size limit on free version

[Bjarnos]

I believe the maximum disk storage for free instances is 2GB, and the unpacked image max 7GB.

Source (7GB): Image download failure for past few days - #2 by nicolas

[Lukasz_Oles]

Hi @EHTASHAM_TOOR, so @Bjarnos is right. The image appears to be too big.
Not the image itself, but one of the layers. It seems that we still need to improve our messaging for this case.

What is your app doing?

[EHTASHAM_TOOR]

Its a python app in which i have a langgraph agent working and using supabase database by a URL. The apps api is provided through FAST API.

Nothing else. Its a simple app. I am also looking for how to reduce the size if it is the problem.

[Bjarnos]

There are a few things you can do to make the size smaller, but I’m not sure which of your images is causing the issue (if you know, feel free to share ).
Here is how you could make your Python installation smaller, if you’re using docker to install it (which I take is the case here, also change 3.13 with your python version):

FROM python:3.13-slim

Also, make sure you’re excluding unnecessary files from the image using a .dockerignore file. Example:

.git
__pycache__
*.pyc
*.log
.env
.venv
node_modules
*.ipynb
.cache

Additionally, when installing Python packages, you can save space by avoiding pip’s cache:

RUN pip install --no-cache-dir -r requirements.txt

Hope this helped!

[EHTASHAM_TOOR]

I was already doing this. and also i removed some files from the project. and also some things added into the docker ignore file. then I saw that image builds up successfully when i was checking the build logs..

then came the instance created. Preparing to start…

and after that instance stopped.. the first picture while reporting the issue.

attaching it again. this image is latest one

[Bjarnos]

It looks like one of the images might still be too large. If you know which image is causing the issue, could you share its name or source? If you’re not sure, please share your Dockerfile so I can take a look at the base image and any other layers that might be contributing to the size. That should help us narrow it down and find a fix.

[Lukasz_Oles]

Yes, the image is too big. The error was confusing and we’ve fixed it.
From the build logs I see that it installs some nvidia-cuda libs. I don’t think you need it in this case.

[EHTASHAM_TOOR]

Yes thank you very much. Now error shows up for image large size if it gets more then the available image size.