fix: update cleanup script to correctly use assumed creds in child accounts

[alharris-at]

Description of changes

There are 2 keys changes in this PR, then a bunch of general cleanup.

1. in configureAws which is called before searching for staleResources in child stacks, we were setting our aws creds to...

accessKeyId: process.env.AWS_ACCESS_KEY_ID,
secretAccessKey: process.env.AWS_SECRET_ACCESS_KEY,
sessionToken: process.env.AWS_SESSION_TOKEN,
...(process.env.AWS_SESSION_TOKEN ? { sessionToken: process.env.AWS_SESSION_TOKEN } : {}),

despite passing in the accessKeyId, etc. So we were never actually using the assumed creds.

2. inside the account->region nested loop we had some code like.

for (const region of AWS_REGIONS_TO_RUN_TESTS) {
  amplifyApps.push(...(await getAmplifyApps(region)));
  stacks.push(...(await getStacks(region)));
}

Because these are parallel network calls, I rewrote the loop to return a list of promises, then yield all promises at once. This could be taken a lot further, since we have nested awaits inside those top-level functions, but that'd require a more significant refactor of the code.

Issue #, if available

N/A, but CI workflows are failing w/ 'Too Many Buckets' or some related error.

Description of how you validated changes

Ran locally w/ my own API key for Circle, and verified the output made sense.

Checklist

• PR description included

By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.

[sundersc]

LGTM.