Full build logs please -

https://bookmarklets.up.railway.app/log-downloader/

Sorry, here is the logs.

You have incorrectly set pip install -r requirements.txt as your build command, that command is automatically ran in the install phase, so running it twice is not the best idea, please remove the build command.

What python version do you use locally?

python version 3.11.3

Add a runtime.txt file to your project with 3.11 in it.

Just 3.11 and nothing else?

Correct.

Ok, I did all the changes and now my full logs are like this

Please attach your requirements.txt file.

Yup!

Please attach your requirements.txt file.

Is there something I can do to fix my issue?

aman, please do not hijack threads, open your own.

syi, that is a 147 dependencies, please work towards removing the deps your app does not use.

But all of those I got from a pip freeze in a free environment where I had to do necessary pip installs of libraries.

But just in case, I did try bringing it down to 30 100% necessary libs but I still get the same deployment fail

Not sure what you mean by "free environment" do you mean venv?

Oh, I meant just a new fresh environment.

Please attach your latest build logs.

Here it is.

Are you sure the versions in your requirements.txt file are the same as what you have installed locally?

yes

Can you try using a newer version of gensim in your requirements.txt?

Ok, I upgraded it to 4.3.2 and is currently building, will let you know if any errors pop up thank you!

The deployment seems to work but now it is crashing at the end. I am assuming it is just a lack of mem because I am using a free version which should be a quick fix that I can do on my own, so I just wanted to honestly say thank you so much for all your help. Appreciate the help Brody .

Awesome, from what I saw about gensim it doesn't seem like something that could be used with only 500MB of memory, but you are also running a development server as far as I can tell? fastapi dev app.py if that is starting a development server, you would be using a lot of unnecessary resources and increased instabilities, I would highly recommend switching to uvicorn.

Would that be an easy change to do?

Well first off, how are you currently defining the start command that Railway is using?

I set a custom start command of fastapi dev app.py

In your service settings on Railway?

yeah

First, make sure you have uvicorn in your requirements.txt.

Then try changing the start command to uvicorn app:app --host 0.0.0.0 --port $PORT

The first app is for app.py and the second app is for what the FastAPI constructor is typically named, change that as needed.

I made the changes and the deployment ended up failing about 5 mins in. Here are the logs:

There's no errors in those logs?

Yeah, I think so. Its just stuck there and I was told that it failed.

Try to redeploy.

Got it

Ok, it finished building but now the deploys logs constantly repeats container event container restart and occasionally shows container died as well, like this:

Starting Container

Jul 10 21:49:11

container event container restart

Jul 10 21:49:39

container event container restart

Jul 10 21:50:10

container event container restart

Jul 10 21:50:43

container event container restart

Jul 10 21:51:12

container event container died

Use the bookmarklet please.

Both logs from build and deployment

To rule out the trial plan being an issue, may I ask you to upgrade to Hobby and redeploy? if you continue to experience such issues, i'll likely get the team involved.

I do believe it is a trial plan issue as I know that running my application is a bit demanding to run. I faced similar problem when trying to deploy using other providers. If I find it impossible to deploy my app for free in other app deploying sites, I will definitely consider trying the upgrade to Hobby. But overall, thank you for all the help, Brody!

No problem, let me know if you continue to experience issues after the upgrade! I'll leave this thread open.