0

I have a Google Cloud function which first creates a Google Room Resource using resources.calendars.insert method from the Google admin sdk, and right after I try to insert an ACL using Acl: insert method from the google calendar api.

Similar to the following code:

const AdminService = google.admin({version: 'directory_v1'});
 try {
    const response = await AdminService.resources.calendars.insert(options); // options omitted
  } catch (error) {
    console.log(`Google Room Resource FAIL`);
    console.error(error.message);
  }

await new Promise((r) => setTimeout(r, 10000));

const CalendarService = google.calendar({version: 'v3'});
 try {
    const res = await CalendarService.acl.insert(option); // options omitted
    console.log(res);
  } catch (error) {
    console.log(error);
    throw new Error(error.message);
  }

As for the authentication, I am using a service account with the correct scopes which impersionates an admin user with the correct permissions. This is how I generate the required JWT token:

const generateJWT = async (scope:string[])=>{
const jwtClient = new google.auth.JWT(
      client_email, // service account
      undefined,
      private_key,
      scope,
      subject // admin user
  );
return jwtClient;
}

In the options parameter for each api call I directly acquire the token for the auth attribute like this:

const option = {
    'calendarId': acl.calendarId,
    'auth': await generateJWT('https://www.googleapis.com/auth/calendar'),
    'resource': {
      'role': acl.role,
      'scope': {
        'type': acl.scopeType,
        'value': acl.scopeValue,
      },
    },
  };

Since I await all api calls, I thought that I will only get the response back when everything is already propagated in Google Workspace but when I do not use the setTimeout in between I always get an Error: Not Found back.

First I had the timeout set to 5 seconds which worked until it didn't so I moved it up to 10 seconds. This worked quite long but now I get again sometimes the Not Found error back...

I don't like the setTimeout hack...and even less if it does not work reliable, so how should I deal with this asynchronous behavior without spinning up any other infrastructure like queues or similar?

zlZimon
  • 2,334
  • 4
  • 21
  • 51

1 Answers1

1

Working with Google Workspace Calendar Resource

As a Super Admin on my organization when creating a Calendar Resource, from the API or the Web interface, it could take up to 24 hours to correctly propagate the information of the Calendar for the organizations, which generally affect the time it would take for any application to gather the ID of the newly created calendar, which could explain why you are increasing the time out.

You have already implemented the await option which is one of the best things you can do. You can also review the option to apply exponential back off to your application or similar to Google App Script a Utitlies.sleep().

There are multiple articles and references on how to utilize it for the retry process needed when the Resource itself has not fully propagated correctly.

You can also review the official Calendar API documentation that suggests that the error "Not Found" is a 404 error:

With a suggested action of reviewing the option to set up exponential backoff to the application.

References:

Ricardo Jose Velasquez Cruz
  • 1,006
  • 4
  • 10
  • Is there anything in the documentation about the propagation time? I could not find anything – zlZimon Sep 29 '22 at 07:12
  • 1
    I know it based on when troubleshooting with the workspace community, and due to the resources generally being created from the Admin console, and being something that affects or applies for the whole organization, the best available documentation from them would be this help center article https://support.google.com/a/answer/7514107 it makes a references to changes being done in the Admin console. – Ricardo Jose Velasquez Cruz Sep 30 '22 at 00:37