Run and Monitor Google Vertex AI Pipelines with Universal Automation Center

Start, monitor, and cache Google Vertex AI pipeline runs directly from a UAC workflow.

We created a new task to run Vertex AI pipelines as part of your Universal Automation Center workflows. Vertex AI pipelines are workflows for executing machine learning pipelines in Google Cloud. Using the Google Vertex AI pipelines task, you can start and monitor Vertex AI pipelines, view every pipeline step, and enable caching so that identical stages of future executions do not have to be redeployed as its output has been cached and can be used directly from the Google bucket. Here you can see the Google Vertex AI pipelines task. You can run and stop pipelines, choose your region, the pipeline template to run, and optional parameters as scripts, or JSON text. Optionally, you can enable caching. Finally, you give a display name to define how the pipeline appears in Google Vertex. Let me run the task. Here in Google Vertex, you can see that the pipeline was started. You can see the display name given in the task. And here, you see the workflow view. All pipeline steps can be viewed in the task instance. Now the task is finished. Here you see the log file with all the steps.

Universal Automation Center now includes a native task for Google Vertex AI Pipelines. Vertex AI Pipelines run machine learning workflows in Google Cloud. With the UAC task, you can start and monitor those pipelines without leaving your automation workflow.

The task gives you full control over every run. Start or stop a pipeline. Choose your region. Pick the pipeline template you want to run. Pass optional parameters as a script or as JSON text. You can also enable caching. When caching is on, identical pipeline stages skip redeployment entirely. UAC pulls the cached output straight from the Google bucket instead. That means faster runs and less wasted compute every time a pipeline repeats a stage it's already completed.

Before you run the task, give the pipeline a display name. That name carries over into Google Vertex, so you always know which UAC task triggered which run.

In the demo, watch the task kick off a Vertex AI pipeline and open directly into the Vertex workflow view. Every pipeline step is visible inside the task instance while it runs. Once the pipeline finishes, the full log file is right there, with every step recorded.

Key Takeaways:

Start and monitor Google Vertex AI pipelines directly from a UAC workflow
Choose your region, pipeline template, and pass parameters as a script or JSON text
Enable caching so repeat runs skip redeployment and pull cached output from your Google bucket
Assign a display name so every pipeline run is traceable back to its UAC task
View every pipeline step, plus the full execution log, inside the task instance

Google Vertex AI Pipeline was rebranded to Gemini Enterprise Agent Platform Pipelines, which is part of the Gemini Enterprise Agent Platform.