In CHAPTER 6.6, Scott says "dynamic content delivery is achieved using HTTP cookies forwarded from your origin"… So, is this the final "result" when I set, for example, S3 objects with a TTL=0 header? Is this the result of a "no-cache" configuration on CloudFront itself? I understand both are valid options, besides choosing to "Forward all cookies to your origin". But notice that Scotts says "FROM" origin, and not "TO" origin. I am a bit confused here… Any guidance? In short, how can I configure CloudFront to serve dynamic content (for instance, from API Gateway) – alternatives and pros/cons.
Dynamic content would be like webpages that were generated by PHP, ASP or a Java process for example…. The URL may be the same, but the page content would be different. I can’t think of an example where you’d have dynamic content coming from an S3 bucket because there’s no server-side processes to render a page dynamically.
Also, for what it’s worth, API Gateway already uses CloudFront behind the scenes so no real need to create a Cloudfront distro in front of an API Gateway deployment for performance reasons. You could use WAF and CloudFront though for extra DDOS protection I suppose.
Hey, Scott, thanks for your answer, but I think you missed my question. I understand the concept of dynamic content and, in fact, it could come, indeed, from S3 as an output repository from a process, e.g., a SPA deploy pipeline saving a new index.html which I don’t want to cache at all. But that’s the point: in this (SPA) case, I change the S3 object metadata informing CloudFront I do not want cache. I would like to know (maybe a Ref.) how a "client cookie" could have the same result on CloudFront cache…
You should be able to find this in the AWS documentation. I found this with a Google search: https://docs.aws.amazon.com/AmazonCloudFront/latest/DeveloperGuide/Cookies.html