Switching from Cloudinary to Cloudflare R2

It is funny that I'm sitting and writing this article in between the time when my AI agent is coding a migration from Cloudinary to Cloudflare R2. Tomorrow is my production release, and I have less than 24 hours now, and I'm in the last few hours to decide whether to switch to a different storage provider. I cannot imagine how crazy it is. But I'm going to be taking this risk. If not for AI, this would not have been possible at all. It would have taken days or weeks to implement this change and do the testing. But I'm going to be doing this in just a few hours. This is something that I like about AI and the way we code and ship.

Source - Gemini

Before I did this activity, I had a detailed plan done with AI, sitting and analysing the code to find out what would be the suitable solution for me. There are so many service providers. The problem with Cloudinary is that it is apparently the costliest among all the other providers. I did a comparison report to find out the available options for us and pick the right one that works for me. When I initially asked the question to AI, it blindly asked me to switch to an AWS S3 bucket. The suggestion was good, but only after continuous discussions was I able to get other options as well. It would have been wrong if I had blindly gone for the AWS S3 bucket.

When it comes to storage, Egress is the most important thing that we have to check. Most of the cloud service providers give free uploads, but when we try to access our own data, they charge us. This is how they operate. Sometimes when we suddenly wake up in the morning, we will be stunned to find out that the cost has increased so much. That can be a problem with the Egress cost, which is where we may not have control. But nowadays, there are alternative service providers that have started giving free Egress.

Currently, switching the storage service in my application is a massive change that I'm making. But considering the advantages it gives, I would like to mitigate the risk before going live so that I don't have to struggle later after the data grows. I wish I had explored this long back to save a lot of time, but not too late even now as AI is helping me out. I'm almost done with this article, and AI has also completed implementing the change in my code. Let me go and review it before taking it live.

If you like what I'm doing on Hive, you can vote me as a witness with the links below.

Vote Hive account@balaz as a Hive Witness

Vote Hive account@kanibot as a Hive Engine Witness