You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I have a small directus instanced hosted on DigitalOcean cloud platform (512mb, 1vCPU, shared)
Very few simple image request are able to completely crash the instance, causing DigitalOcean to restart it, without any logs other then was terminated
While I understand that a small instance like that might not be very stable, I think just a few images by a single user should not be able to completely crash the system causing it to restart
To Reproduce
I think the simplest way to reproduce it is:
Create the 512mb, 1vCPU shared directus instance on DO
Using the directus API: Upload around 10 jpg images (around 4000x6000px, 3mb each)
Simply open the directus app and navigate to the folder you uploaded the files to
This should cause the app to generate thumbnails the first time you open it up. The directus instance will most likely crash (I have a 100% success rate with 10 fresh images 🙂)
Directus Version
v10.10.4
Hosting Strategy
Self-Hosted (Docker Image)
The text was updated successfully, but these errors were encountered:
Image transformation is a very CPU/Memory intensive process. The images you mention "around 4000x6000px, 3mb each" are almost the default MAX transform size of 6000x6000px and doing multiple of those may require more than the minimum specs.
Can you try lowering the ASSETS_TRANSFORM_MAX_CONCURRENT config to see if it results in a more stable experience? And perhaps increasing the ASSETS_TRANSFORM_TIMEOUT could help too.
Even having ASSETS_TRANSFORM_MAX_CONCURRENT at 2 crashes the server. It crashes immediately much earlier before any timeout comes into play, so I don't think it has got to do something with ASSETS_TRANSFORM_TIMEOUT .
Scaling the app to 1GB RAM instead of 512mb immediately solves the issue and it perfectly handles concurrent transformations (at least from the one user I am)
I am closing this issue since I unfortunately can only conclude that this was a resource limitation somewhere when handling the images (be it not enough RAM or CPU). This is unfortunately something we cannot currently fix at the minimum required resources https://docs.directus.io/self-hosted/docker-guide.html#requirements
Definitely feel free to create a new issue if it turns out to related to the specific images used or something along those lines 😄
Describe the Bug
I have a small directus instanced hosted on DigitalOcean cloud platform (512mb, 1vCPU, shared)
Very few simple image request are able to completely crash the instance, causing DigitalOcean to restart it, without any logs other then
was terminated
While I understand that a small instance like that might not be very stable, I think just a few images by a single user should not be able to completely crash the system causing it to restart
To Reproduce
I think the simplest way to reproduce it is:
This should cause the app to generate thumbnails the first time you open it up. The directus instance will most likely crash (I have a 100% success rate with 10 fresh images 🙂)
Directus Version
v10.10.4
Hosting Strategy
Self-Hosted (Docker Image)
The text was updated successfully, but these errors were encountered: