- Newest
- Most votes
- Most comments
Which cache policy is applied to the cache behaviour leading to the S3 origin in CloudFront? Importantly, is it set to cache content or to disable caching? If caching is disabled, every request will go to the S3 origin, both increasing your costs and potentially leading to intermittent issues if the request limits for S3's internal partitions in your bucket are exceeded.
Have you got S3 server access logs enabled for your bucket? The request rate throttling would show there as SlowDown errors (https://docs.aws.amazon.com/AmazonS3/latest/userguide/ErrorBestPractices.html#UsingErrorsSlowDown). Also, if the failures are caused by some other issue with S3, the logs would reveal specifics in most cases.
You can also estimate whether rate limiting is involved by looking at the metrics for your CloudFront distribution, and/or by enabling request metrics (https://docs.aws.amazon.com/AmazonS3/latest/userguide/cloudwatch-monitoring.html) for your bucket. They will show separately from CloudFront the numbers of requests and their distribution across different types of requests (like read and write) for your bucket.
GIven that this is an intermittent issue I recommend that you create a support case and get the support team to dig into why you're seeing failures. Make sure you include request ids as well as times/dates.
@Aditi, did you ever resolve the above issue? If so, what was the resolution steps that were followed, as I'm currently experiencing similar behaviour across my platforms.
