No error details when trying to deploy a google cloud function

[arsanjea]

While trying to deploy a google cloud function with pulumi, I get very little information as to what went wrong which leads to hours of troubleshooting efforts. Example: It took me a lot of troubleshooting to figure out why pulumi wasn't able to deploy the following piece of code:

const functionNodejs = new gcp.cloudfunctions.Function("hello", {
  environmentVariables: {
    NODE_ENV: 'production',
  },
  sourceArchiveBucket: bucket.name,
  runtime: "nodejs8",
  sourceArchiveObject: bucketObjectNodejs.name,
  entryPoint: "hello",
  triggerHttp: true,
  availableMemoryMb: 512,
  timeout: 540,
});

Pulumi's output was:

Diagnostics:
  pulumi:pulumi:Stack (test):
    error: update failed

  gcp:cloudfunctions:Function (hello):
    error: Plan apply failed: 1 error occurred:
        * updating urn:pulumi:test/function:Function::hello: Error while updating cloudfunction configuration: googleapi: Error 400: The request has errors, badRequest

After hours of troubleshooting I figured out that NODE_ENV was not environment variable a user is allowed to set.

I was able to identify this issue by running: gcloud functions deploy hello --update-env-vars NODE_ENV=production --trigger-http

Which gives the - much more helpful - following output:

dist zeus$ gcloud functions deploy hello --update-env-vars NODE_ENV=production --trigger-http --runtime nodejs8
ERROR: (gcloud.functions.deploy) ResponseError: status=[400], code=[Bad Request], message=[The request has errors
Problems:
environment variable name NODE_ENV is reserved by the system: it cannot be set by users
]

Would it possible for pulumi to capture this output?

[CyrusNajmabadi]

That's very unfortunate. We can see what is possible, but we may be very limited ultimately by what the underlying terraform provider does. In order to improve this, it might necessitate a change at their layer to some collect and pass along more information when their call into the googleapi fails.