Monitoring With Flower

Accessing the Flower Web Interface

Once the installation of Blurit is complete, Flower's web interface is accessible at:

http://<your-server-address>/flower

Use the username and password configured during the installation of Blurit to log in.

Overview of the Interface

The Flower interface provides an intuitive way to monitor and manage Celery tasks and workers in real-time. Below is a detailed breakdown of the key sections and columns in the interface.

Dashboard: Monitoring Workers Overview

The Dashboard tab in Flower provides an overview of all Celery workers currently running. This page allows you to quickly monitor worker states and performance.

Column Descriptions

1. Worker Name

   • Displays the unique name of each worker. Typically, the name is prefixed with celery@ followed by a unique identifier (often based on the hostname).

   • Example: celery@18e5d5189283.

2. Status

   • Indicates the current status of the worker.

   • Online: The worker is active and communicating with the broker.

   • Offline: The worker is not responding or has been stopped.

3. Active

   • The number of tasks currently being processed by this worker.

   • A high value here relative to the worker’s capacity could indicate overload.

4. Processed

   • The total number of tasks processed by the worker since it started.

5. Failed

   • The total number of tasks that have failed on this worker.

   • A high number may indicate recurring issues with specific tasks or code.

6. Succeeded

   • The total number of tasks successfully executed by the worker.

7. Retried

   • The number of tasks that were retried after encountering a temporary error.

8. Load Average

   • The system's average load (CPU, memory) for the container or machine running the worker.

   • Format: 1 min, 5 min, 15 min.

   • Example: 0.19, 0.08, 0.08 indicates a low load across all three timeframes.

Available Actions

• Search Bar: Use the search bar in the top-right corner to filter workers by name or other attributes. This is helpful when managing many workers.

• Refresh: Click the Refresh button to update the data displayed on the dashboard.

Tasks: Monitoring and Managing Individual Tasks

The Tasks tab in Flower allows you to monitor the state and details of individual Celery tasks in real-time. This tab is especially useful for diagnosing specific issues or analyzing task performance.

Column Descriptions

1. Name

   • The name of the task, as defined in your Celery code.

   • Example: blurit.image.

   • This typically corresponds to a function or method in your application.

2. UUID

   • A unique identifier for each task, useful for debugging or referencing specific tasks.

   • Example: c2e583bc-1dc9-4193-9df5-fe412aa0b180.

3. State

   • The current state of the task.

   • Common states include:

     • PENDING: The task is waiting to be executed.

     • STARTED: The task is currently running.

     • SUCCESS: The task has completed successfully.

     • FAILURE: The task execution failed.

   • Example in the screenshot: SUCCESS.

4. args

   • The positional arguments passed to the task when it was executed.

   • Example: () (no positional arguments in this case).

5. kwargs

   • The keyword arguments (named parameters) passed to the task.

   • Example:

     jsonCopier le code{
         "input_image": {"protocol": "file", "path": "input/1732010550292-480.jpg"},
         "output_image": {"protocol": "file", "path": "output/878ac326-2b6c-482e-9888-e4346d285ad0.jpg"},
         "lifespan": 3600
     }

6. Result

   • The result returned by the task after execution.

   • Example (summary):

     jsonCopier le code{
         "input_image": {"protocol": "file", "path": "input/1732010550292-480.jpg"},
         "output_image": {"protocol": "file", "path": "output/878ac326-2b6c-482e-9888-e4346d285ad0.jpg"},
         "chronos": {"total": 2.219586, "childs": {...}}
     }

7. Received

   • The timestamp when the task was received by the worker.

   • Example: 2024-11-19 10:02:30.472.

8. Started

   • The timestamp when the task started execution.

   • Example: 2024-11-19 10:02:30.478.

9. Runtime

   • The total time the task took to execute, in seconds.

   • Example: 2.733.

10. Worker

    • The Celery worker that executed the task.

    • Example: celery@83b122840b58.

Usage Tips

1. Search for Tasks

   • Use the search bar to filter tasks by UUID, Name, or State. This is especially helpful when dealing with large numbers of tasks.

2. Diagnose Failures

   • For tasks in the FAILURE state, the kwargs and Result columns provide valuable details about what went wrong.

3. Performance Optimization

   • Use the Runtime column to identify tasks that take a long time to execute, which may need optimization.

4. Input and Output Validation

   • The kwargs and Result columns are also useful for verifying the input and output data of specific tasks.