Handle azure-cns pod errors or mark node as not-ready when node is initializing/deallocation

**Component (Azure NPM or Azure CNI):**
CNI/CNS & containerd componets - could be change in AKS node lifecycle

**Describe in detail the feature/behavior/change you'd like to see:**
customers see intermittent issues in their infra during node deallocation (starting, initializing or restart) , azure-cns pods have errors in being ready or starting up due to file errors related to cni.
Similar issues filed before - https://github.com/Azure/azure-container-networking/issues/2999, https://github.com/Azure/AKS/issues/4342
Customer wants a feature where the Node is marked as Not ready till the cns-pods are Running & in good state.
AKS & ACN needs to come up with a feature & decide on which component this can be handled in future so the intermittent issues are not seen in azure-cns pods.

**Orchestrator(e.g. Kubernetes, Docker):**
AKS/K8s

**Operating System (Linux/Windows):**
Both

**Anything else you would like to add:**
ICM link to follow for all details - https://portal.microsofticm.com/imp/v5/incidents/details/695606517/summary

Start by reviewing the related issues #2999 and AKS #4342, then trace how Azure CNI/CNS and AKS/Kubernetes handle node initialization and deallocation on Linux and Windows. The issue names no repository file or test; done requires an agreed component and feature that prevents intermittent azure-cns startup or readiness errors by keeping the node NotReady until CNS is healthy.

Written by the indexing model from the issue text.