[BUG] OME seems to scale down Deployments to 1 replica temporarily whenever it changes something

[fzyzcjy]

What happened?

After changing sth, e.g. a InferenceService/Runtime definition like how the probe should be like, etc, I see the auto-generated Deployment go through:

• Before: e.g. Deployment spec replica = 100
• After: Deployment spec replica = 1 for a while (this looks weird)
• After a while: Deployment spec replica = 100 (this is normal)

I have not digged this problem yet, thus not very sure whether it is ome bug or not, but given my helm chart is nothing but simply change the InferenceService/runtime. If you cannot reproduce feel free to ping me and I can try to ablate into a min sample.

What did you expect to happen?

How can we reproduce it (as minimally and precisely as possible)?

# Example YAML configuration that reproduces the issue

Anything else we need to know?

Environment

• OME version: 89f217e
• Kubernetes version (use kubectl version):
• Cloud provider or hardware configuration:
• OS (e.g., from /etc/os-release):
• Runtime (SGLang, vLLM, etc.) and version:
• Model being served (if applicable):
• Install method (Helm, kubectl, etc.):