Can we trigger a vertex AI pipeline using github trigger?

Google Cloud Build with AI Custom ML & MLOps
1

I have created a vertex AI pipeline using KFP in workbench and it is working fine. Can I use the same code to deploy the vertex AI pipeline using github action when there is any change in the code?

I have tried and I am able to copy the pipeline_definition.json to GCS and able to access the google big query tables. However the below code is not working,

pipeline.run(
service_account = SERVICE_ACCOUNT
)

I have ensured that the SERVICE_ACCOUNT has all the necessary permissions. Any support would be much appreciated.

Thanks,
Prakash

1 Like
2

If you’re able to copy the pipeline definition to Google Cloud Storage (GCS) and access Google BigQuery tables successfully but encountering issues when running the pipeline using pipeline.run(service_account=SERVICE_ACCOUNT), there might be several reasons why this is happening:

Ensure that you’re passing the correct service account credentials to the pipeline.run() method. Verify that SERVICE_ACCOUNT variable contains the correct service account JSON key file or the appropriate credentials needed for authentication.

Wrap the pipeline.run() method in a try-except block to catch any exceptions that might be raised during execution. This will help in identifying the specific error that’s occurring.

Show More

try:

Show More

pipeline.run(service_account=SERVICE_ACCOUNT)

Show More

except Exception as e:

Show More

print(“Error running pipeline:”, e)

Make sure the environment in which you’re running the pipeline has all necessary dependencies installed and properly configured.

1 Like