New RunLLM response generated:
>>>*# Question*
š I am trying to make a Flyte workflow that does the following:
⢠Runs a task
⢠Launches a dynamic workflow via a for loop
⢠This calls another dynamic workflow, also with a for loop
⢠(*) For each of these calls, call a task
⢠This task has several additional tasks called from within it
⢠One of these tasks, midway through, is another dynamic task
I noticed a few things and had some questions:
1. When I create this workflow initially, the UI does not show any parallelism after the second dynamic task. That is, the third dynamic task does not appear
2. The UI also does not break apart the task marked with a (*) into its sub-tasks, making the UI more obscure to what is happening
Any insight here appreciated!
# Answer
To address your concerns:
1.
UI Parallelism Visibility: The Flyte UI may not display nested dynamic workflows beyond a certain level due to limitations in how dynamic workflows are visualized. This is a known issue where the UI does not always reflect the full hierarchy of nested dynamic workflows. According to
Flyte GitHub Issue #4466, there are ongoing discussions about improving error visibility and UI representation for dynamic workflows.
2.
Task Breakdown in UI: The Flyte UI might not break down tasks into their sub-tasks if they are encapsulated within a dynamic workflow or if the task is complex. This can make it challenging to visualize the workflow's execution path. As per
Flyte GitHub Issue #5164, there are known issues with dynamic tasks not being fully represented in the UI, which can lead to confusion.
Would you like more detailed information on how to potentially work around these UI limitations?
Sources:
-
Flyte GitHub Issue #4466
-
Flyte GitHub Issue #5164
Slack Link
https://.slack.com/archives/CP2HDHKE1/p1729727655.810019 š 0 button š 0 button š”
Hint: Mention @runllm in the thread for followups.