I’m not 100% sure about your first question, but I don’t think there an abstraction / setting for this cc
@thankful-minister-83577.
You could use FlyteRemote to check if a different workflow is running from inside of a dynamic task and then decide not to continue.
For question two I believe your cluster will have
• quota set at the project level
• resource request and limits at the task level
• Only a certain number of machines
• Certain type of machines
There are a lot of limits to worry and out and different issues, especially if you are going to hit different limits. One solution is limiting the parallelism of a workflow
# for a specific workflow
LaunchPlan.get_or_create(
name="cpu_intensive_wf",
workflow=cpu_intensive_wf,
max_parallelism=10,
)
# in helm
flyteadmin:
maxParallelism: 25
see
docs
Yee will have a better answer late! He is core dev of Flyte.