Playground should be triggered to create Services and then Pods once the model is created #109
Conversation
InftyAI-Agent
added
needs-triage
|
/cc @kerthcet |
|
/kind bug |
InftyAI-Agent
added
bug
carlory
force-pushed
the
PlaygroundReconciler
branch
2 times, most recently
from
d265621 to
95258d1
Compare
| }, | ||
| checkPlayground: func(ctx context.Context, k8sClient client.Client, playground *inferenceapi.Playground) { | ||
| validation.ValidatePlayground(ctx, k8sClient, playground) | ||
| gomega.Consistently(func() error { |
There was a problem hiding this comment.
I'm not sure if no condition is expected.
There was a problem hiding this comment.
Yes for now, but this can be optimized like pending for model creation in the follow up if you're interested or together with this PR, whatever you like.
So when there's no condition with Playground or model is not created, Playground is Pending, but with different reasons, one is Waiting for inferenceService ready, another is Waiting for model to be created.
There was a problem hiding this comment.
n the follow up if you're interested or together with this PR, whatever you like.
I will do it in another PR.
carlory
force-pushed
the
PlaygroundReconciler
branch
from
95258d1 to
61c31f6
Compare
| }, | ||
| checkPlayground: func(ctx context.Context, k8sClient client.Client, playground *inferenceapi.Playground) { | ||
| validation.ValidatePlayground(ctx, k8sClient, playground) | ||
| gomega.Consistently(func() error { |
There was a problem hiding this comment.
Yes for now, but this can be optimized like pending for model creation in the follow up if you're interested or together with this PR, whatever you like.
So when there's no condition with Playground or model is not created, Playground is Pending, but with different reasons, one is Waiting for inferenceService ready, another is Waiting for model to be created.
carlory
force-pushed
the
PlaygroundReconciler
branch
from
61c31f6 to
ad9973f
Compare
…the model is created
carlory
force-pushed
the
PlaygroundReconciler
branch
from
ad9973f to
ba653c6
Compare
|
/lgtm Thanks! |
InftyAI-Agent
added
lgtm
3 participants
What this PR does / why we need it
Let's image there's no Model in the cluster, I created a Playground, because no Model exists, inference Pods will not be created. However, if I created the corresponding Model, the Playground should be triggered to create Services and then Pods. Right now, this is not true.
The solution is quite simple, make Playground watch for model creation events.
Which issue(s) this PR fixes
Fixes #92
Special notes for your reviewer
Does this PR introduce a user-facing change?