Q: We are using HTTP Direct with IONs in cloud and having trouble with larger responses throwing 500s. Conversation with AI suggests that this could be jetty g-zipping the response and the API gateway not being able to handle the chunked, g-zipped response. We haven’t figured out how to enable logs to see the true problem source. Does anyone have any experience with behavior like this?
Are you running into the 30s API Gateway limitation? https://stackoverflow.com/questions/54299958/how-can-i-set-the-aws-api-gateway-timeout-higher-than-30-seconds/54300708#54300708
This Cloudwatch metric should tell you:
No, the handler returns in 2 secs in the logs so it's size, not time which is triggering the problem. I'm gonna try to enable jetty logging to see if that exposes the truth
We saw similar behaviour when not using HTTP Direct but the size threshold was larger.
I’ve been trying to figure out how to disable GZIP in the Jetty container run in the cloud, but I’m failing hard on that challenge.