Flower does not display full Celery task input/output when values are large or nested

[1hemmem]

Description / Context:

Flower is an excellent dashboard for monitoring Celery, but I'm running into an issue where it truncates or hides full task input arguments and/or return values, especially when they're large dictionaries, long nested data, or deeply nested JSON. This makes debugging and auditing difficult.

Steps to Reproduce:

1. Define a Celery task that receives or returns a large or deeply nested dict/JSON.
2. Launch Flower and invoke the task with that payload.
3. Observe the Task Details view—look at the “Arguments” and “Return Value” fields.

[JoseSilva22]

I had the same problem
The only way I managed to solve this issue was to use the flag result_extended = True as they say on this page: stack overflow discussion

And then go get the args there with something like:

def get_task_by_id_backend(events, task_id):

    connection = psycopg2.connect(
        database="celeryresults",
        user="postgres",
        password="postgres",
        host="db_celery",
        port=5432,
    )

    with connection.cursor(cursor_factory=RealDictCursor) as cur:
        cur.execute(f"SELECT * from celery_taskmeta where task_id = '{task_id}';")
        row = cur.fetchone()

    task = events.state.tasks.get(task_id)

    try:
        task.args = pickle.loads(row["args"].tobytes())[0]
    except Exception as e:
        print(e)
    return task