Proper error reporting

[albertodiazdorado]

Although this action works nicely when things work out correctly, the error reporting when things go wrong is lacking. In fact, the error reporting is so poor that, when I need to debug a certain class of errors, I need to add a step to my github actions issuing the command aws cloudformation deploy just so that I can debug the problem.

I can provide a concrete example. My deployment just failed, and aws-cloudformation-github-deploy issues the following error:

Error: 1 validation error detected: Value 'AWSTemplateFormatVersion: 2010-09-09
Transform: AWS::Serverless-2016-10-31
(...)
(the beginning of the template follows)

With that kind of error I am clueless about how to fix my deployment. When I use the AWS CLI I obtain:

Templates with a size greater than 51,200 bytes must be deployed via an S3 Bucket. Please add the --s3-bucket parameter to your command. The local template will be copied to that S3 bucket and then deployed.
Error: Process completed with exit code 255.

---

I can provide a single example right now, but I have found myself in this situation multiple times. Thus:

1. Am I doing something wrong? Do we have a documented/undocumented --debug flag that I should be using?
2. If not: are you interested in reworking the error reporting? Should I try to open a PR?

Honestly! I am thinking of using some sort of debug to print this! Some help to make life easier would be amazing or if any body has any better ideas?

[waggonerjake]

One thing that I have used in the past to help catch errors is the tool cfn-lint. It can be used to lint your templates for common issues or discrepancies. It doesn't help with deployment errors, but it's a good way to catch errors prior to deployment.

[albertodiazdorado]

@waggonerjake cfn-lint is a good tool to use, that however provides utility that is orthogonal that what I am asking for here.

This issue is not about preventing errors, but about understanding the error when they happen.

[LoganB99]

Same thing happened to us today with the 51,200 bytes. I wish this was fixed