Storage usage due to imgix

Hi there

I have my own S3 bucket and we upload and download data to it via API connections from bubble.

We are using imgix to resize, crop and watermark images used in our app. The resulting files, once processed by imgix, turn up in the bubble s3 bucket.

In our workflow, once processed by imgix, we download the resulting file, and upload it to our own S3 bucket.

I dId not consider that on-the-fly image rendering by imgix would leave the file in our bubble s3 bucket forever and at the moment my bubble file manager has 23,000 files in it taking up over 10GB of space. Obviously I have also the same data in my own S3.

I don’t know if we could, once used, find a file that has popped up in the file manager, and delete it via the API or somesuch. We are on the professional plan and our final application will have the better part of a terabyte of images in storage. There is no way that we could possibly keep on top of deleting the images generated one at a time in that file manager page.

Has anyone solved this problem before?


OK further to the above I seem to be able to do THIS in my workflows


Which in theory I could call with 's file URL and clean up as I go, so I don’t make the problem any worse

So the cleanup…

Does anyone know if you can query down the API into the file manager to get your list of files in storage (which then you can use an action in a workflow to delete)

I know I can copy and paste from the file manager screen into excel. If I click “next 50” 20 times to get a thousand records on screen at once, and I can write a macro to extract the URLs into a list, and I can import that list and build a workflow on it, so in theory I can make this not too painful a cleanup.

Users in my app will upload files to us all the time. Which I am about to build to use the bubble uploader dropzone, which results in a list of files. It’s going to be very important for me, to destroy those files once they have been uploaded if they are not needed any more.

Lessons learned!

1 Like

hi richard, I probably to use the same way later. will try to implement first then see if that works for me. will keep in touch