CHORE(resources): set memory limits equal to memory requests

Align memory limits with memory requests for guaranteed QoS class.
- prometheus, thanos (query, storegateway, compactor)
- alertmanager, tempo, goldilocks (dashboard, controller)
- node-exporter, opentelemetry-collector, vpa, kube-state-metrics

alertmanager/helm-values.yaml

@@ -26,7 +26,7 @@ resources:
     cpu: 15m
     memory: 100Mi
   limits:
-    memory: 150Mi
+    memory: 100Mi
 
 # Disable default config - use secret instead
 config:

goldilocks/helm-values.yaml

@@ -11,7 +11,7 @@ dashboard:
       cpu: 15m
       memory: 100Mi
     limits:
-      memory: 150Mi
+      memory: 100Mi
 
   service:
     type: ClusterIP
@@ -54,7 +54,7 @@ controller:
       cpu: 15m
       memory: 100Mi
     limits:
-      memory: 150Mi
+      memory: 100Mi
 
   # Enable VPA recommendations for all namespaces
   # Set to false to only monitor namespaces with the label: goldilocks.fairwinds.com/enabled=true

kube-state-metrics/helm-values.yaml

@@ -12,7 +12,7 @@ resources:
     cpu: 15m
     memory: 100Mi
   limits:
-    memory: 150Mi
+    memory: 100Mi
 
 # Schedule on control-plane node
 nodeSelector:

loki/helm-values.yaml

@@ -20,9 +20,9 @@ loki:
       ruler: loki
       admin: loki
 
-  # Single binary mode with 2 replicas
+  # Single binary mode with 1 replica
   commonConfig:
-    replication_factor: 2
+    replication_factor: 1
 
   # Schema config
   schemaConfig:
@@ -49,18 +49,16 @@ loki:
 deploymentMode: SingleBinary
 
 singleBinary:
-  replicas: 2
+  replicas: 1
   persistence:
     enabled: false
-  affinity:
-    podAntiAffinity:
-      preferredDuringSchedulingIgnoredDuringExecution:
-      - weight: 100
-        podAffinityTerm:
-          labelSelector:
-            matchLabels:
-              app.kubernetes.io/name: loki
-          topologyKey: kubernetes.io/hostname
+  # Run on master node for stability (consistent with Thanos/Tempo)
+  tolerations:
+    - key: node-role.kubernetes.io/control-plane
+      operator: Exists
+      effect: NoSchedule
+  nodeSelector:
+    node-role.kubernetes.io/control-plane: "true"
   extraEnvFrom:
     - secretRef:
         name: minio-s3-credentials

node-exporter/helm-values.yaml

@@ -11,7 +11,7 @@ resources:
     cpu: 15m
     memory: 64Mi
   limits:
-    memory: 96Mi
+    memory: 64Mi
 
 service:
   type: ClusterIP

opentelemetry-collector/helm-values.yaml

@@ -28,7 +28,7 @@ resources:
     cpu: 25m
     memory: 64Mi
   limits:
-    memory: 256Mi
+    memory: 64Mi
 
 # =============================================================================
 # Tolerations (run on all nodes including master)

prometheus/helm-values.yaml

@@ -86,7 +86,7 @@ prometheus:
         cpu: 50m
         memory: 512Mi
       limits:
-        memory: 768Mi
+        memory: 512Mi
     
     # ServiceMonitor 자동 발견 - 모든 ServiceMonitor 선택
     serviceMonitorSelectorNilUsesHelmValues: false

tempo/helm-values.yaml

@@ -23,7 +23,7 @@ resources:
     cpu: 50m
     memory: 128Mi
   limits:
-    memory: 512Mi
+    memory: 128Mi
 
 # =============================================================================
 # Tempo Configuration

thanos/helm-values.yaml

@@ -56,7 +56,7 @@ query:
       cpu: 15m
       memory: 128Mi
     limits:
-      memory: 256Mi
+      memory: 128Mi
 
 # =============================================================================
 # Query Frontend - Caching layer for Query (optional, disabled for small cluster)
@@ -84,7 +84,7 @@ storegateway:
       cpu: 15m
       memory: 128Mi
     limits:
-      memory: 512Mi
+      memory: 128Mi
 
   persistence:
     enabled: true
@@ -115,7 +115,7 @@ compactor:
       cpu: 15m
       memory: 128Mi
     limits:
-      memory: 512Mi
+      memory: 128Mi
 
   persistence:
     enabled: true

vpa/helm-values.yaml

@@ -11,7 +11,7 @@ recommender:
       cpu: 15m
       memory: 128Mi
     limits:
-      memory: 192Mi
+      memory: 128Mi
 
   # Schedule on control-plane node
   nodeSelector: