Updated deployment

Author: nullchimp

.github/workflows/deploy-azure.yml

@@ -154,25 +154,54 @@ jobs:
           kubectl get pods
           kubectl get services
           
-          echo "Waiting for Memgraph pod to be ready (may take up to 5 minutes)..."
-          if ! kubectl wait --for=condition=ready pod -l app=memgraph --timeout=5m; then
+          echo "Waiting for Memgraph pod to be ready (may take up to 10 minutes)..."
+          # First, check if there are any existing failed pods and delete them
+          FAILED_PODS=$(kubectl get pods -l app=memgraph -o jsonpath="{.items[?(@.status.phase=='Failed' || @.status.containerStatuses[0].state.waiting.reason=='CrashLoopBackOff')].metadata.name}")
+          if [[ -n "$FAILED_PODS" ]]; then
+            echo "Found failed pods, cleaning up before proceeding:"
+            echo "$FAILED_PODS" | tr ' ' '\n' | while read pod; do
+              echo "Deleting failed pod: $pod"
+              kubectl delete pod $pod
+            done
+            echo "Waiting for cleanup to complete..."
+            sleep 30
+          fi
+          
+          # Wait for pod to be ready with increased timeout and check interval
+          if ! kubectl wait --for=condition=ready pod -l app=memgraph --timeout=10m; then
             echo "Error: Memgraph pod did not become ready within the timeout period."
             echo "Checking Memgraph pod logs:"
             POD_NAME=$(kubectl get pods -l app=memgraph -o jsonpath="{.items[0].metadata.name}")
-            kubectl logs $POD_NAME
+            
+            echo "Pod details:"
             kubectl describe pod $POD_NAME
-            echo "Deployment verification failed!"
+            
+            echo "Pod logs:"
+            kubectl logs $POD_NAME --tail=100
+            
+            echo "Checking node resource utilization:"
+            kubectl describe nodes
+            
+            echo "Deployment verification failed! Please check the Kubernetes cluster's resources and capacity."
             exit 1
           fi
           
           echo "Memgraph deployment successful!"
           
           # Wait for the LoadBalancer service to get an external IP
           echo "Waiting for LoadBalancer to get external IP..."
-          for i in {1..30}; do
+          for i in {1..45}; do
             EXTERNAL_IP=$(kubectl get service memgraph -o jsonpath='{.status.loadBalancer.ingress[0].ip}')
             if [[ -n "$EXTERNAL_IP" ]]; then
               echo "Memgraph is accessible at: ${EXTERNAL_IP}:7687"
+              
+              # Add a simple connectivity check
+              echo "Performing basic connectivity test..."
+              if nc -z -w 5 ${EXTERNAL_IP} 7687; then
+                echo "Successfully connected to Memgraph on port 7687!"
+              else
+                echo "Warning: Could not connect to Memgraph on port 7687. The service may not be fully ready yet."
+              fi
               break
             fi
             echo "Waiting for external IP (attempt $i)..."
@@ -181,4 +210,6 @@ jobs:
           
           if [[ -z "$EXTERNAL_IP" ]]; then
             echo "Warning: Could not obtain external IP for Memgraph service within timeout."
+            echo "Checking LoadBalancer service status:"
+            kubectl describe service memgraph
           fi

infra/azure/main.tf

@@ -62,7 +62,12 @@ resource "azurerm_kubernetes_cluster" "ai_agent" {
   default_node_pool {
     name       = "default"
     node_count = 1
-    vm_size    = "Standard_D2_v2"
+    vm_size    = "Standard_D4_v2"  # Increased from D2 to D4 for more resources
+    
+    # Enable auto-scaling for better resource management
+    enable_auto_scaling = true
+    min_count           = 1
+    max_count           = 3
   }
 
   identity {

infra/k8s/memgraph.yaml

@@ -28,6 +28,21 @@ spec:
         command: ["sysctl", "-w", "vm.max_map_count=262144"]
         securityContext:
           privileged: true
+      # Add node affinity to schedule on nodes with sufficient resources
+      affinity:
+        nodeAffinity:
+          preferredDuringSchedulingIgnoredDuringExecution:
+          - weight: 100
+            preference:
+              matchExpressions:
+              - key: beta.kubernetes.io/instance-type
+                operator: In
+                values:
+                - Standard_D4_v2
+      tolerations:
+      - key: "node.kubernetes.io/memory-pressure"
+        operator: "Exists"
+        effect: "NoSchedule"
       containers:
       - name: memgraph
         image: memgraph/memgraph-mage:latest
@@ -56,11 +71,11 @@ spec:
           mountPath: /etc/memgraph
         resources:
           requests:
+            memory: "512Mi"
+            cpu: "250m"
+          limits:
             memory: "1Gi"
             cpu: "500m"
-          limits:
-            memory: "2Gi"
-            cpu: "1"
         livenessProbe:
           httpGet:
             path: /api/v1/storage/status