Hello everyone, I am trying out Flyte, so what I have observed is that node id are typically of type n0, n1, n1-n1, etc but today when I tried nested condition the node id was a random string second nested node, for example

fyamacey

, why is it so? is there any logic behind this or is this completely random? and how this id is generated?

@icy-breakfast-53872 is this a dynamic workflow? There are a couple of references that suggest this is by design, mainly to reduce pressure on etcd capacity, where wf execution status is persisted. Right now it seems to be that for deeply nested workflows (with >20 nodes) Flyte moves from sequential node IDs to a hash (ref). Unfortunately this field is not configurable yet. Does that help?

We want to create unique ids that are deterministic- these hare hashed ids

okay, Thank you for the explanation @average-finland-92144 @freezing-airport-6809 this makes sense, but our nodes were not more than 10 but yes the workflow had nested conditions, so that can be the reason?

@icy-breakfast-53872 correcting here:

Right now it seems to be that for deeply nested workflows (with >20 nodes) Flyte moves from sequential node IDs to a hash

It's not the number of nodes but the lenght of the nodeId, measured in number of characters, what triggers the encoder to go to hashed IDs. So it might be the case that even with 10 nodes and nested conditions, the hashing algorithm is activated

okay got it, Thank you @average-finland-92144 .