incus-apply does not wait even for kernel to be started in vm created from custom image

Hello,

I create vm1 alpine1 from a standard `images:alpine/3.23/cloud` or `images:alpine/3.24/cloud`, both work fine, `incus-apply` even waits for `cloud-init` to complete and indicates all phases correctly.

Then I replace a kernel by `apk add` with my custom built kernel, reboot vm manually by `incus restart`, it works fine, boots correctly.

Now I clean out `cloud-init` configs and publish created vm1 as a new custom image and create a new vm2 from the image using `incus-apply` again. And now `incus-apply` does not wait enough even for vm2 to start, far from even trying to start `cloud-init`.

I checked by `incus console` command and discovered that kernel is being started more later than `incus-apply` already exited.

How can I fix this? Please let me know? How to force `incus-apply` to wait in vm2 kernel being started at least?

Thinking about moving to `Terraform` / `OpenTofu`.