fix(AWS API Gateway): Proper stage resolution for custom resource

[pgrzesik]

After recent changes the bug surfaced where the stage for custom resource needed for API Gateway logs did not take into account setting from provider.stage

Closes: #9240, #9268

[codecov]

Codecov Report

Merging #9277 (a78005f) into master (dac06c8) will decrease coverage by 0.00%.
The diff coverage is 100.00%.

❗ Current head a78005f differs from pull request most recent head 072452f. Consider uploading reports for the commit 072452f to get more accurate results

@@            Coverage Diff             @@
##           master    #9277      +/-   ##
==========================================
- Coverage   86.95%   86.95%   -0.01%     
==========================================
  Files         315      315              
  Lines       11736    11735       -1     
==========================================
- Hits        10205    10204       -1     
  Misses       1531     1531              

Impacted Files	Coverage Δ
lib/plugins/aws/customResources/index.js	98.57% <100.00%> (-0.03%)	⬇️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update dac06c8...072452f. Read the comment docs.

[medikoo]

Looks great 👍

[chekalsky]

@pgrzesik @medikoo Hi! Does this PR breaks BC? One of my lambdas wasn't deploying with "Modifying service token is not allowed.." error and I suppose that problem is here (probably name change of a function). That alone doesn't really bother me as I can remove lambda and recreate it later, but this also somehow messed with CORS headers in either CloudFront responses, either in Api Gateway, and these headers aren't being send.

[pgrzesik]

Hello @chekalsky - it should fix a previously missed bug - stage wasn't resolved correctly and was working "by chance" and it should not introduce any breaking changes but it's definitely possible. Could you provide a reproducible example of a bug you're experiencing so we could take a deeper look into it? Thanks in advance

[Frikitrok]

@pgrzesik hey, i am trying to upgrade my SLS infrastructure from 2.32.1 to latest 2.68.0 and can see just same error @chekalsky described. So i have different lambda handlers for cognitoUserPool and S3 events, all failing on "sls deploy" with error "An error occurred: CognitoCustomUserPool - Modifying service token is not allowed..". I would assume this is because SLS defining custom event resources in another way, respecting stage variable. Any suggestion how to migrate to latest version without downtime?

[pgrzesik]

Hey @Frikitrok - sorry you've run into trouble. Did you add any other modifications to your config during the update or you just upgraded the Framework version and attempted a deployment?

[Frikitrok]

@pgrzesik If it helps, i've removed serverless-pseudo-parameters plugin because now it supported on SLS vanilla, and added new one serverless-prune-plugin.

[Frikitrok]

@pgrzesik ^

[pgrzesik]

Are you able to share more about your configuration so I could try to reproduce it on my side?