Hi - I've got a few questions about the speed of the API to help me dial in my expectations: I'm seeing some pretty slow performance (15 mins to upload 96 ~200KB files from an AWS lambda instance with an eager transform), but I'm not 100% sure yet if its Cloudinary or my misusing guzzle promises thats causing the slowdown (its all happening via a job queue on AWS lambda so debugging is a bit of a pain).
Out of curiosity, could you let me know the expected processing & response time of the upload API for an image upload with eager transform (e.g. just resizing from 4096px square to 2048px)?
Is passing s3 paths faster in anyway than direct uploading? (as I mentioned, I'm on AWS lambda so its all via AWS network regardless). When I am transferring directly to/from s3 the same files take <1s.
Would it be advisable to upload the file without an eager transform, then grab it w/transform later? I assume there might be some copying etc happening in the background that the client also ends up waiting on when doing eager transforms.
I was also thinking maybe I'm hitting some kind of soft-API limit (like only processing one image every 5 seconds or similar). Are there any limits beyond the 500 images per hour I should be aware of?
(also on an unrelated note, I'd like to join your startup programme, but couldn't get a response via the form. I'm building a SaaS from scratch to generate 3D models for AR e-commerce, and I'd like to fully integrate cloudinary, especially for your USDZ/GLTF support down the line, but as you'd expect the budget's quite tight while just starting out!
Please sign in to leave a comment.