AWSPinpointProvider error when executing GraphQL operations on project without Analytics category

There seems to be changes introduced in aws-amplify that causes AWS PinPoint code to execute when doing GraphQL operations even though Analytics is not added to the Amplify project.

The error happens when executing GraphQL operations server side (TypeScript)

Error: No credentials, applicationId or region

    at AWSPinpointProvider.<anonymous> (C:\Users\meiri\dev\betterbook\betterbook-backend\node_modules\@aws-amplify\analytics\src\Providers\AWSPinpointProvider.ts:159:36)
    at step (C:\Users\meiri\dev\betterbook\betterbook-backend\node_modules\@aws-amplify\analytics\lib\Providers\AWSPinpointProvider.js:40:23)
    at Object.next (C:\Users\meiri\dev\betterbook\betterbook-backend\node_modules\@aws-amplify\analytics\lib\Providers\AWSPinpointProvider.js:21:53)
    at fulfilled (C:\Users\meiri\dev\betterbook\betterbook-backend\node_modules\@aws-amplify\analytics\lib\Providers\AWSPinpointProvider.js:12:58)
    at process._tickCallback (internal/process/next_tick.js:68:7)

This started happening after upgrading my project to the latest versions, although I’m not sure which package causes the issue. (This also relates to #3486 that happens after the version upgrade, I added the workaround polyfill as suggested there and now this happens.

Environment:

node 10.16.0 amplify 1.7.4

"dependencies": {
    "@aws/dynamodb-batch-iterator": "^0.7.1",
    "@sentry/browser": "^5.4.0",
    "@sentry/node": "4.4.2",
    "@types/aws-lambda": "^8.10.27",
    "@types/babel__traverse": "^7.0.7",
    "@types/yargs": "^13.0.0",
    "amazon-cognito-auth-js": "^1.3.2",
    "amazon-cognito-identity-js": "^3.0.12",
    "apollo-cache-inmemory": "^1.6.2",
    "aws-amplify": "^1.1.29",
    "aws-appsync": "^1.8.0",
    "aws-lambda-multipart-parser": "^0.1.3",
    "aws-sdk": "^2.474.0",
    "aws4": "^1.8.0",
    "axios": "^0.18.1",
    "axios-es6": "^0.11.1",
    "bcrypt-nodejs": "^0.0.3",
    "blueimp-md5": "^2.10.0",
    "body-parser": "^1.19.0",
    "cryptr": "^4.0.2",
    "errorhandler": "^1.5.1",
    "es6-promise": "^4.2.8",
    "form-data": "^2.3.2",
    "google-libphonenumber": "^3.2.2",
    "graphql-tag": "^2.10.1",
    "handlebars": "^4.1.2",
    "isomorphic-fetch": "^2.2.1",
    "jsonwebtoken": "^8.5.1",
    "jwk-to-pem": "^2.0.1",
    "lodash": "^4.17.11",
    "node-cmd": "^3.0.0",
    "node-fetch": "^2.6.0",
    "node-jose": "^1.1.3",
    "qs": "^6.7.0",
    "request": "^2.88.0",
    "tabletop": "^1.5.2",
    "uuid": "^3.3.2",
    "ws": "^7.0.0"
  },

[ghost]

When using AWS Amplify CLI (or MobileHub), the configuration is contained in the auto-generated aws-exports.js.

Seeing these PinPoint issues re-appearing from time to time, maybe Amplify-CLI could add the Analytics: { disabled: true } iff the CLI has not added the Analytics service.

In the meantime,

import awsExports from './aws-exports';
Amplify.configure({
   ...awsExports,
   Analytics: { 
       disabled: true
   }
});

[aehmt]

  Amplify.configure({
    Auth: {
      region: config.region,
      userPoolId: config.userPoolId,
      userPoolWebClientId: config.userPoolWebClientId,
      authenticationFlowType: 'USER_PASSWORD_AUTH'
    },
    Analytics: {
      disabled: true
    }
  })

Fixes the issue.