fix: Resolve errors for vm resource not found

[tangyouzzz]

Fixes #525

Description
https://learn.microsoft.com/en-us/azure/azure-resource-manager/troubleshooting/error-not-found?tabs=bicep#symptoms
When using the GET method to obtain a VM, the Error Code returned when the VM does not exist can be either NotFound or ResourceNotFound.

{"level":"ERROR","time":"2024-12-24T08:08:51.172Z","logger":"controller","message":"Reconciler error","commit":"56aa2e0-dirty","controller":"node.termination","controllerGroup":"","controllerKind":"Node","Node":{"name":"aks-cloud-fip-mcu-arm64-mkr2f"},"namespace":"","name":"aks-cloud-fip-mcu-arm64-mkr2f","reconcileID":"2ee3a64a-083f-4743-84fa-661820165f2c","error":"ensuring instance termination, getting cloudprovider instance, getting instance, failed to get VM instance, GET https://management.azure.com/subscriptions/xxxx/resourceGroups/xxxx/providers/Microsoft.Compute/virtualMachines/aks-cloud-fip-mcu-arm64-mkr2f\n--------------------------------------------------------------------------------\nRESPONSE 404: 404 Not Found\nERROR CODE: NotFound\n--------------------------------------------------------------------------------\n{\n  \"error\": {\n    \"code\": \"NotFound\",\n    \"message\": \"The entity was not found in this Azure location.\"\n  }\n}\n--------------------------------------------------------------------------------\n"}

In IsNotFoundErr, only the situation of ResourceNotFound will be judged, which will cause the node to remain in the notready state and cannot be deleted.

func IsNotFoundErr(err error) bool {
	azErr := IsResponseError(err)
	return azErr != nil && azErr.ErrorCode == ResourceNotFound
}

How was this change tested?

• Using az vm simulate-eviction to simulate spot node release

Does this change impact docs?

• Yes, PR includes docs updates
• Yes, issue opened: #
• No

Release Note

[tangyouzzz]

@microsoft-github-policy-service agree

[Bryce-Soghigian]

LGTM @tallaxes take a look for second approval?

[Bryce-Soghigian]

https://github.com/Azure/azure-sdk-for-go-extensions/blob/main/pkg/errors/armerrors.go#L34 can we also add this change here too?

[coveralls]

Pull Request Test Coverage Report for Build 12678035442

Details

• 4 of 4 (100.0%) changed or added relevant lines in 2 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage increased (+0.001%) to 95.425%

---

Totals
Change from base Build 12661198057:	0.001%
Covered Lines:	48136
Relevant Lines:	50444

---

💛 - Coveralls

[tallaxes]

LGTM - and thank you @tangyouzzz for the fix! Some possible improvements:

• Checking for azErr.StatusCode == http.StatusNotFound instead should catch both
• Same logic should be applied deleting other resources (like NIC); other resource providers are very likely to exhibit the same pattern
• A helper method would be better to avoid code duplication (and yes, would be good to update azure-sdk-for-go-extensions)

[tallaxes]

Merging since this fixes an actual problem, but let's follow-up on feedback