The runs tab provides a visual overview of your workflow's execution timeline using a waterfall chart. It allows you to monitor performance, debug issues, and analyze how individual steps in your workflow are executing.
You can access this view by opening your workflow without selecting a specific step and clicking the "runs" tab in the overview panel.
Header statistics
At the top of the tab, you will see a summary of your workflow performance:
Runs: The total count of persistent (non-ephemeral) runs.
Run: The timestamp of the currently selected run, along with its status:
Succeeded: All steps completed successfully.
Failed: One or more steps encountered errors.
Running: The execution is currently in progress.
Duration: The total time elapsed from the start of the first step to the completion of the last.
Steps: The number of steps that executed during this run.
The waterfall chart
The main visualization displays each step's execution as a horizontal bar, allowing you to see how long each part of your workflow takes.
Queue time (lighter color): The duration the step waited before it began executing.
Run time (darker color): The actual time the step took to complete.
Error indicator (red circle with X): Indicates that a step failed.
Steps are ordered chronologically, with the top bar representing the first step to start. Hovering over any bar will reveal a detailed tooltip containing the step name, queue time, actual run time, and total duration.
Key features for analysis
Step highlighting: Hovering over a bar highlights the corresponding step in your workflow map, and vice-versa.
Click to navigate: Clicking any bar takes you directly to that specific step's detailed view.
Real-time updates: The chart updates automatically as new steps start or complete, allowing you to monitor active runs.
Performance insights: Use the chart to identify bottlenecks, find steps with long queue times, and verify that parallelization is working as expected.
Navigating runs
If your workflow has multiple runs, you can switch between them using the run picker dropdown.
This menu displays all persistent runs, their status, and relative timestamps, making it easy to compare performance across different executions.
Jump between executions
When viewing a specific step execution, arrow buttons let you navigate to adjacent runs without returning to the list:
Previous and next run. The down and up arrows jump to the immediately older or newer execution in the list.
Hover to reveal more targets. Hold your pointer over an arrow for a moment, and the control expands to show two additional jump options:
Previous or next success (green checkmark): Jump to the nearest execution that completed successfully.
Previous or next failure (red X): Jump to the nearest execution that failed. These buttons search the full retained history of failures, not just the loaded window, so you can step through every error without scrolling.
These controls appear when a step has more than one execution, making it easy to compare runs, investigate a sequence of failures, or confirm that a fix resolved an issue.