CloudFront 5xxErrorRate without technical evidences

Question

I'm receiving a lot of error messages that suggest the resources associated with my CloudFront distributions are/were outage. Nevertheless, I cannot view any messages/logs associated with unavailable resources throughout AWS Health Dashboard.

The error message I've receiving looks like as informed below:

> The AWS CloudFront instance XYZ is returning 5XX errors for 70.5882% of requests, putting it into a state of error. You should check that CloudFront is up and configured properly for your environment.
> If this is a side-effect of an AWS outage (eg. S3 is unavailable and your content is stored there), you should see an alert on the "AWS Service Health" EventSource soon.
> This condition has been ongoing since 2024-03-04 09:30:00 BRT, or for 0h 24m."

How can I troubleshoot this particular scenario?

Accepted Answer

There could be multiple reasons.

Did you check your content is available in S3 ? ( if thats the origin ? ).
Any policy/permission change on s3 ? ( in that the origin? ).

Next step would be have a look at this documentation to correlate the issue.

https://docs.aws.amazon.com/AmazonCloudFront/latest/DeveloperGuide/troubleshooting-response-errors.html

You may need to enable cloudfront logs to dig deep into actual error code.
https://docs.aws.amazon.com/athena/latest/ug/cloudfront-logs.html

Answer

My hunch would be something behind your api gateway is throwing the 500 errors.

I would look at your application logs and like Muhammad said enable cloudfront logs.

CloudFront 5xxErrorRate without technical evidences

Relevanter Inhalt