Topology Lifecycle State and Probe Integration

[bayars]

Summary

This MR adds a full lifecycle state machine to the Topology CRD, giving operators real-time visibility into where a lab is in its deployment lifecycle. It's improving the integration of the clab in the different applications.

It also exposes per-node probe statuses for fine-grained health observability.

The new STATE column is visible directly in kubectl get topologies.

The state machine general design is like this:

| State | What it means |
|-------|---------------|
| `deploying` | Nodes are starting up, not ready yet |
| `deployfailed` | A node crashed before the lab ever came up |
| `running` | All nodes are healthy |
| `degraded` | Was running, but a node went down |
| `destroying` | Delete was requested, waiting ~5s before cleanup |
| `destroyfailed` | Deletion got stuck, needs manual investigation |

I tried to document the states in the code. I will write a docs to explain all details in a separate documentaiton for lifecycle state mechanism, and probe integrations.