dialogflow.v3.Agent needs the start_flow in update calls

[dedsm]

What happened?

After the resource is created, any modification will fail because the API call requires the agent start playbook to be specified.

Example

  google-native:dialogflow/v3:Agent (diag-agent):
    error: error sending request: googleapi: Error 400: com.google.apps.framework.request.BadRequestException: Agent start playbook or flow should be specified.: 

Output of pulumi about

CLI          
Version      3.96.2                       
Go Version   go1.21.4                                                                                                                                                                                                                                                            
Go Compiler  gc                                                                                                                                                                                                                                                                  

Plugins
NAME           VERSION
gcp            7.2.2
google-native  0.32.0
python         unknown

Host     
OS       debian
Version  12.2
Arch     x86_64

This project is written in python: executable='/usr/local/bin/python3' version='3.9.18'

Dependencies:
NAME                  VERSION
pip                   23.0.1
pulumi-gcp            7.2.2
pulumi-google-native  0.32.0
setuptools            58.1.0

Pulumi locates its logs in /tmp by default

Additional context

same behavior happens with dialogflow.v3beta1.Agent

Contributing

Vote on this issue by adding a 👍 reaction.
To contribute a fix for this issue, leave a comment (and link to your pull request, if you've opened one already).

[mjeffryes]

Thank you for reporting this @dedsm! If you need a workaround in the short term, I'd recommend using the gcp.diagflow.Agent from the GCP provider.

[dedsm]

hi @mjeffryes thanks for the quick reply, sadly I'm using the integration with Vertex AI Search and Conversation and other things that are not available in the terraform module, if you have an idea on another alternative I'm all eyes :)