Add node watcher functionality and test case

[hlts2]

This PR includes the implementation of a watcher that periodically checks the status of nodes in a Kubernetes cluster, specifically focusing on the availability of GPUs and the readiness of the nodes.

The watcher works as follows:

it fetches the list of nodes that match the specified selector.
It then checks if the node meets the desired GPU count and whether the node is in a ready state. If either condition is not met, the node is considered unready.
If a node is found to be unready or doesn't have the expected GPU count, the watcher attempts to reboot the node by invoking a reboot on the corresponding instance in Civo.
This functionality ensures that nodes are automatically monitored and rebooted if necessary to meet the specified conditions, maintaining optimal performance and readiness.

⚠️ It also includes the necessary test cases to validate the functionality. I am currently checking the GPU part, so some corrections might be made later, but the implementation and tests are complete for now.

[hlts2]

@jokestax @johndietz

Thank you for your review! I will merge this PR 👍