Navigating Through Airflow UI
This article guides you through the Apache Airflow UI, explaining how to navigate key features like DAGs, task status, logs, and monitoring tools for efficient workflow management.
Over the years, Airflow's UI has been functional but often felt outdated and cluttered, making it difficult to navigate through its many components. With the release of Airflow 3, the UI and overall user experience have been significantly improved, offering a cleaner layout, better organisation, and more intuitive interactions.
In this article, I’ll walk you through the updated Airflow 3 UI, highlighting key components and sharing practical tips for tasks like monitoring logs, understanding schedules, and working with data assets more effectively.
Airflow UI
Users can access to Airflow UI from http://localhost:8080/ address. The UI has several tabs for different requirements and each one of them can be accessed from the left section. At this article we are going to through the tabs and check functionalities.
Home Tab
Home section is the default section we meet when we access to the Airflow UI. This tab is quite important for the overall health of the DAGs and Tasks in a glimpse.
At the image above, we have three main stats sections; DAGs, Tasks, and Assets. These sections give us aggregated statistics about DAGs, Tasks, and Assets. The DAGs and Tasks sections are customised by the Time filters which have the last 24 hours as default variable. We can select different timeframes and the stats are going to be recomputed accordingly.
Once critical statistics, and often misinterpreted, on this UI is the Active DAGs. Active DAGs doesn’t represent the amount of DAGs in your DAG folder. It represents the ones are not turned off from the DAGs section on the left bar. So basically your total number of DAGs counted as a sum of active and inactive DAGs and this stats section only highlights active ones. For the total sum we need to navigate to the DAGs section.
Dags Tab
Dags tab is the section where we can monitor all the DAGs in Airflow regardless they are active or not. It has three sub sections as DAGs, Runs, and Task Instances.
Dags
As seen from the image above, the yellow vertical highlights the DAGs and their names. The Airflow DAG names or ids should be unique so we should expect no recurring names on this UI. Under the DAG name we can see the schedule behaviour. In Airflow 3 we can have different types of schedules such as time based or asset triggered. Time based schedules can be identified easily with a calendar component.
The green vertical highlights the Latest Run of the specific DAG. The sign next to the timestamp indicates the status of the run such as success, failure, etc.
The blue vertical highlights the Next Run of the specific DAG. This column is populated if the DAG is a time based scheduled. If the DAG is asset triggered, then we should not expect any pre-assigned timestamp here as we might not know when the asset is going to be generated.
Runs
Runs section is the detailed view of the DAG Runs. It hold various important information about the DAG and its associated Run.
Run Type: This column indicates what type of schedule or trigger mechanism was used to run the DAG
Duration: This column indicates how long the latest run took.
Dag Version: DAGs are live code pieces that evolve over time. As they evolve, they get incremental versions. Especially when we are migrating between the versions, using this column is crucial how was the behaviour of the specific DAG version.
Task Instances
Similar to DAG Runs, this section hold detailed information about the Tasks that are underlying the DAGs. While having similar components with the other sections, Tasks Instances have also some unique information as well.
Task ID: The unique task that does something under the given DAG.
Operator: The operator that was used to perform the task. On the image above we can see two types of operators: native operators or decorated operators.
Assets
In Airflow 3, assets represent named units of data—like files, tables, or any resource identified by a URI—that can be produced or consumed by tasks. They are a major upgrade from the older dataset concept, enabling more intelligent, event-driven scheduling.
When a task outputs an asset (using @asset decorators or the outlets parameter), Airflow records this event, and downstream DAGs that depend on that asset can be triggered automatically. This shifts workflows from time-based execution to data-aware orchestration, meaning DAGs can react the moment fresh data becomes available.
Assets also support advanced scheduling logic using expressions like asset1 & asset2, letting you wait for multiple assets before triggering a run. You can even integrate asset events with external sources like message queues.
Browse
The browse section has two parts: Audit log and XComs.
Audit Logs serve the purpose of tracking and recording key user actions and system events within the Airflow environment. This includes operations like triggering DAGs, modifying variables, changing connections, or editing configurations through the UI or API. The goal is to provide visibility into who did what and when, which is crucial for accountability, debugging, and maintaining a secure, well-governed system. Audit Logs help teams monitor changes, trace issues, and comply with internal policies or external regulations.
XComs (short for Cross-Communication) are used to allow tasks to share small pieces of data with each other during a DAG run. This is helpful when one task needs to pass results—like a file path, ID, or status—to another downstream task. XComs enable this communication behind the scenes, making task coordination more dynamic and data-aware. They’re especially useful in complex workflows where the output of one step directly influences the next.
Admin
The Browse section has five parts: Connections, Variables, Pools, Providers, and Plugins.
Connections
Here you configure and manage the credentials and settings required for Airflow to interact with external systems like databases, cloud services, and APIs. This centralised interface lets you add, edit, or test integration points for your workflow tasks.
Variables
This section stores global key‑value pairs that your DAGs can reference at runtime. Variables are great for managing dynamic inputs—like file paths or thresholds—without altering your code.
Pools
Pools help limit the number of concurrent tasks that can run in shared environments, such as databases or APIs. Here you can create named pools and define how many slots each pool offers to prevent overwhelming critical resources.
Providers
This area lists all installed provider packages (e.g., AWS, GCP, Snowflake) along with their versions. It lets you verify which integrations are available, ensuring compatibility with external systems.
Plugins
Use this section to view any custom operators, sensors, hooks, or UI components that have been added to extend Airflow's base functionality. It’s a quick way to confirm your plugin modules are loading correctly.
Security
The Security section has five parts: Users, Roles, Actions, Resources, and Permissions.
Users
This section lists every account with access to Airflow, including their login name, email, and assigned roles. You can add new users, update existing profiles, or deactivate accounts.
Roles
Here, you define or edit roles—such as Admin, Op, User, and Viewer—and assign which actions each role can perform. Roles streamline permission management by grouping capabilities together and applying them to users.
Actions
This area catalogues the different operations available in Airflow, like "can_read" or "can_edit". Actions describe what can be done, apart from where it applies, and form the basic unit of behaviour in the RBAC model.
Resources
Resources are the UI objects or endpoints you control access to—such as DAGs, task instances, connections, or XComs. This section lets you view which parts of Airflow are protected by permissions.
Permissions
Permissions link Actions to Resources within a role. For example, a permission might grant the “Op” role the ability to "can_edit" the "Connection" resource. This is where access rules are formally defined and enforced.
Conclusion
Exploring the Airflow 3 UI reveals how much thought and refinement has gone into improving the user experience. From a more intuitive layout to deeper insights through features like Audit Logs, XComs, and Assets, the interface now makes it easier than ever to monitor, manage, and troubleshoot workflows. Whether you're building DAGs, managing access, or inspecting task runs, the new UI supports a smoother and more efficient workflow. With these enhancements, Airflow 3 isn't just more powerful under the hood—it's also more accessible on the surface.