When a scheduled DAG runs repeatedly, the grid view provides a time-oriented record of its executions and task outcomes. Columns represent individual DAG runs, while rows represent tasks across those runs. section 2.4 section 1.2.4
Each task cell shows the result for one task instance in one run. Selecting a cell opens task details and logs, making the view useful for tracing failures across a pipeline rather than inspecting only its structure. section 1.2.4
Use the grid view to handle an unsuccessful task:
- Find the failed task in the affected DAG run; dependent tasks may appear as upstream_failed and will not run. section 2.5
- Open the task instance and inspect its logs or stack trace to identify causes such as connectivity problems, code errors, or unavailable services. section 2.5
- After correcting the problem, clear the failed task instance instead of restarting successful work. section 2.5
- Choose whether to clear only that task or include upstream, downstream, or corresponding tasks in another run. section 2.5
Scheduled pipeline recovery
A daily DAG contains a task that downloads data and a dependent notification task. If the download loses network access, the grid view marks the download task failed and the notification task upstream_failed. Inspect the download logs, restore connectivity, clear the failed task and necessary downstream state, and let Airflow rerun the unfinished portion. section 2.4 section 2.5
