[ML] Trained Model: Fix start deployment with ML autoscaling and 0 active nodes

[darnautov]

Summary

During my testing, I used the current user with all required privileges but failed to notice that, after switching to the internalkibana_system user, it lacked the manage_autoscaling privilege required for the GET /_autoscaling/policy API.

As a result, the isMlAutoscalingEnabled flag, which we rely on in the Start Deployment modal, was always set to false. This caused a bug in scenarios with zero active ML nodes, where falling back to deriving available processors from ML limits was not possible.

You can check the created deployment, it correctly identifies ML autoscaling:

Checklist

Check the PR satisfies following conditions.

• [ ] Unit or functional tests were updated or added to match the most common scenarios

[elasticmachine]

Pinging @elastic/ml-ui (:ml)

[elasticmachine]

:green_heart: Build Succeeded

• Buildkite Build
• Commit: 2ec92186bc7bef257a2da43c6028411f2e1a9146
• Cloud Deployment

Metrics [docs]

✅ unchanged

History

cc @darnautov