tirkarthi commented on PR #40755:
URL: https://github.com/apache/airflow/pull/40755#issuecomment-2236582957
It looks easier to implement switching between line and bar chart. In the
latest commit I have added a checkbox similar to "show landing time" in run
duration without persistence to make it easier to play around locally if
someone is interested. I also wanted to see how the chart is for dags with lot
of tasks where many of our dags have minimum 10 tasks and below is a sample dag
to generate random sleep tasks.
Line chart to me looks unusable when there are lot of data points that
finish around same time but useful to see trends as noted with fewer number of
tasks with varying time. Bar chart also looks useful where user can click on
the bar to get to task detail like line chart and looks better in some cases.
With a lot of tasks the bars/lines can get crammed and might be harder to
select the exact one without unselecting few tasks in the legend.
Few thoughts and questions :
* Lot of tasks means legend can be larger and possibly collides with y axis.
I tried various options in echarts like padding, nameGap but couldn't get it
working.
* Task groups don't filter down to further tasks as separate bars.
* I have seen the chart go negative while running might be some issue in
calculation of duration for running tasks.
* Maybe keep the checkbox so that user can select the best representation as
per the situation.
* Tooltip could probably display only the relevant datapoint instead of time
for everything. I am not sure how to select the right option for that in
echarts.
```python
from datetime import datetime, timedelta
from airflow import DAG
from airflow.decorators import task, task_group
with DAG(
dag_id="all_task_duration",
start_date=datetime(2024, 7, 1),
catchup=True,
schedule_interval="@daily",
) as dag:
def base(i, j):
import random, time
duration = random.randrange(i, j)
for index in range(duration):
time.sleep(1)
print(index)
@task
def extract():
base(10, 20)
@task
def load():
base(20, 30)
@task
def transform():
base(2, 5)
@task
def transform_1():
base(2, 5)
@task
def transform_2():
base(2, 5)
@task
def transform_3():
base(2, 5)
@task
def transform_4():
base(2, 5)
@task
def transform_5():
base(2, 5)
@task
def transform_6():
base(2, 5)
@task
def transform_7():
base(2, 5)
@task
def transform_8():
base(2, 5)
@task
def transform_9():
base(2, 5)
@task
def transform_10():
base(2, 5)
@task
def transform_11():
base(2, 5)
@task_group(group_id="my_task_group")
def tg1():
@task
def tg11():
base(11, 20)
@task
def tg12():
base(20, 30)
tg11() >> tg12()
(
extract()
>> load()
>> tg1()
>> transform()
>> transform_1()
>> transform_2()
>> transform_3()
>> transform_4()
>> transform_5()
>> transform_6()
>> transform_7()
>> transform_8()
>> transform_9()
>> transform_10()
>> transform_11()
)
```
Screenshots :
Lot of tasks with same duration where bar chart looks better than line chart
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
