Memory Flask Hobby Plan
Project ID: 638d95a1-0f0d-4801-96d4-e68d6701d8a6
I also attach a screenshot of the deploy log.
I am completely new to app deployment and this is my first try. I trained a model with tensorflow, and I want to deploy it with flask and run with gunicorn. I tested the app locally and it works. To load the model, I need in total about 800 MB of RAM. When deploying on railway with the trial plan, it crashes (SIGKILL), probably due to memory. Makes sense, since I had only 512MB RAM on the trial plan. I upgraded to hobby, where I should have 8GB. I redeployed, also tried creating a new project and redeploying several times. It still dies with the same error. The memory print messages and railway metrics indicate that the deployment dies right before reaching 512MB. How can I ensure the project is using the 8GB of my Hobby plan?
5 Replies
Project ID:
638d95a1-0f0d-4801-96d4-e68d6701d8a6
a full redeploy should move your service onto the higher limits, can you show me your memory metrics
Solution
that does indeed look like a hard cap is being hit
unfortunately I don't have much to say, maybe try this https://bookmarklets.up.railway.app/service-redeploy/