Memory sizing is reset to 256MB on every deploy

Dear Support Team,

I’m experiencing two issues with my Phoenix application:

  1. Out of Memory (OOM) errors:
  • Initially encountered OOM errors with 256MB RAM
  • Scaled to 512MB, which resolved the issue temporarily
  1. Persistent scaling problems:
  • After each deployment, the app reverts to 256MB
  • This causes deployment failures and OOM errors to recur

Additional concerns:

  • Will frequent scaling changes impact billing?
  • Why does a relatively simple CRUD API require so much memory?

I’d appreciate your assistance in:

  1. Investigating why the scaling doesn’t persist across deployments
  2. Advising on optimal memory allocation for a basic Phoenix CRUD API
  3. Clarifying any potential billing implications

Thank you for your help in resolving these issues.

Best regards,
Benjamin

I had to add this to my fly.toml to make it works — sc2x comes with 512MB by default.

[[vm]]
size = 'shared-cpu-2x'

But I find it weird that these OOM issues suddenly appeared for a lot of people using the 256MB lately. Clearly something has changed. Yet no real answers :frowning: :

It seems strange that a 2 users CRUD API under development in Phoenix can’t run on 256MB… But I’m no expert so who knows :man_shrugging:

From App not working to Questions / Help

This topic was automatically closed 2 days after the last reply. New replies are no longer allowed.