Amplify build failing with Resource XXXXX must be in ARN format or "*"

**Describe the bug**
Amplify console build is failing with the following error after adding a storage category and creating a trigger for it.

```
2021-03-17T15:26:54.616Z [WARNING]: Error updating cloudformation stack
2021-03-17T15:26:59.620Z [INFO]: Following resources failed
Resource Name: ampli-batc-1OX21FKOVDKP8 (AWS::IAM::Policy)
Event Type: update
Reason: Resource storagebatchJobStateArn must be in ARN format or "*". (Service: AmazonIdentityManagement; Status Code: 400; Error Code: MalformedPolicyDocument; Request ID: fa959ce9-327b-4dd7-bec2-a276456e0133; Proxy: null)
URL: XXXXX
Resource Name: userManager (AWS::Lambda::Function)
Event Type: update
Reason: Resource update cancelled
URL: XXXXX
```

Steps to reproduce the behavior:
1. Add storage
2. Configure trigger
3. push
4. build fails

**Expected behavior**
A clear and concise description of what you expected to happen.
The build should work.

**Desktop (please complete the following information):**
- OS: MacOS
- Browser: Chrome
- Version: Version 89.0.4389.82 (Official Build) (x86_64)

**Additional context**
This issue happened locally first, and I fixed it by adding the storage resource to the 'dependsOn' array in the Trigger definition in the amplify/backend/amplify-meta.json file (as described here: https://github.com/aws-amplify/amplify-cli/issues/5333). I've tried removing the amplify-meta.json file from gitignore and pushing again, but the same results.

Start with the storage trigger configuration and amplify/backend/amplify-meta.json, especially the Trigger definition and its dependsOn array. Reproduce an Amplify build after adding storage and a trigger, then inspect the generated IAM policy and CloudFormation update. Done means the build succeeds without the resource ARN format error.

Written by the indexing model from the issue text.