Updated delete script to be more careful

Author: nullchimp

.github/workflows/deploy-azure.yml

@@ -51,8 +51,11 @@ jobs:
           ./import_resources.sh
 
       - name: Terraform Apply
+        timeout-minutes: 120  # Increase timeout to 2 hours for AKS node pool operations
         run: |
           cd infra/azure
+          # Set TF_PLUGIN_TIMEOUT to increase the timeout for Azure provider operations
+          export TF_PLUGIN_TIMEOUT=120m
           terraform apply -auto-approve \
             -var="environment=${{ env.ENVIRONMENT }}" \
             -var="memgraph_username=${{ secrets.MEMGRAPH_USERNAME }}" \

infra/azure/main.tf

@@ -161,6 +161,15 @@ resource "null_resource" "cleanup_unmanaged_resources" {
 set -e
 echo "Starting cleanup of unmanaged resources in ${data.azurerm_resource_group.github.name} resource group..."
 
+# Check if there are any ongoing operations on the AKS cluster
+AKS_NAME="aks-ai-agent-${var.environment}"
+AKS_PROVISIONING_STATE=$(az aks show --name $AKS_NAME --resource-group ${data.azurerm_resource_group.github.name} --query "provisioningState" -o tsv 2>/dev/null || echo "NotFound")
+
+if [[ "$AKS_PROVISIONING_STATE" != "Succeeded" && "$AKS_PROVISIONING_STATE" != "NotFound" ]]; then
+  echo "AKS cluster is currently in $AKS_PROVISIONING_STATE state. Skipping cleanup to avoid conflicts."
+  exit 0
+fi
+
 # Get all resources in the resource group
 ALL_RESOURCES=$(az resource list --resource-group ${data.azurerm_resource_group.github.name} --query "[].id" -o tsv)