A data analyst is monitoring a newly deployed Dataflow batch pipeline in the Google Cloud console to verify its health and check for potential bottlenecks. While viewing the pipeline details, they need to inspect the status of individual processing stages (such as Running, Succeeded, or Failed) and evaluate step-level performance details. Which feature in the Dataflow job UI should the analyst use to inspect individual stage statuses and their associated step metrics?

Click on individual stage nodes within the Dataflow execution graph to view the stage summary and step metrics in the side panel.: ### **What the Dataflow execution graph is:** The **Dataflow execution graph** (also known as the job graph) is the visual representation of a pipeline's execution steps within the Google Cloud console. In Dataflow, Google Cloud optimizes Apache Beam pipelines by fusing compatible transforms into execution **stages**. The console visually represents these stages as connected nodes that show execution status in real time. ### **How it meets the requirements:** - **Inspect stage statuses (✓):** Each node in the execution graph displays an icon and color indicating its current execution state—such as **Not started**, **Running**, **Succeeded**, **Failed**, or **Stopped**. - **Evaluate step-level performance (✓):** Selecting an individual stage node opens a dedicated side panel containing detailed stage summaries, including **Elapsed time**, **Fused steps**, **Input collections**, **Output collections**, and custom metrics. - **Identify performance bottlenecks (✓):** Analysts can immediately see which specific stage is currently executing, how long it has taken relative to other stages, and whether worker resource usage or data skew is causing delays. ### **Key advantages and capabilities:** - **Visual progress tracking:** Provides an immediate visual snapshot of where data is flowing and which parts of the pipeline are active or finished. - **Fusing transparency:** Shows exactly which transforms from the Apache Beam code were fused together into a single execution stage. - **Targeted troubleshooting:** Allows users to isolate errors directly to a specific stage and filter logs specifically for that transform step. - **Real-time metric breakdown:** Displays input/output record counts and throughput rates on selected nodes during execution. ### **Why this is optimal:** Interacting directly with the **execution graph** is the standard, native way to drill down into stage-level execution within Dataflow. It links high-level pipeline flow directly to granular stage metrics and filtered worker logs without requiring custom scripts or navigating away from the job dashboard.