PERF(observability): remove CPU limits for stability

- Remove CPU limits from all observability components
- Prevents CPU throttling issues across monitoring stack

alertmanager/helm-values.yaml

@@ -21,13 +21,12 @@ affinity:
 persistence:
   enabled: false
 
-# Resource settings (VPA lowerBound/target)
+# Resource settings (no CPU limit for stability)
 resources:
   requests:
     cpu: 15m
     memory: 100Mi
   limits:
-    cpu: 15m
     memory: 100Mi
 
 # Disable default config - use secret instead

blackbox-exporter/helm-values.yaml

@@ -5,13 +5,12 @@ fullnameOverride: blackbox-exporter
 
 replicas: 1
 
-# Resource settings (VPA lowerBound/upperBound)
+# Resource settings (no CPU limit for stability)
 resources:
   requests:
     cpu: 15m
     memory: 100Mi
   limits:
-    cpu: 32m
     memory: 100Mi
 
 config:

goldilocks/helm-values.yaml

@@ -6,13 +6,12 @@ dashboard:
   enabled: true
   replicaCount: 1
 
-  # Resource settings (VPA lowerBound/upperBound)
+  # Resource settings (no CPU limit for stability)
   resources:
     requests:
       cpu: 15m
       memory: 100Mi
     limits:
-      cpu: 15m
       memory: 100Mi
 
   service:
@@ -51,13 +50,12 @@ controller:
   enabled: true
   replicaCount: 1
 
-  # Resource settings (VPA lowerBound/upperBound)
+  # Resource settings (no CPU limit for stability)
   resources:
     requests:
       cpu: 15m
       memory: 100Mi
     limits:
-      cpu: 25m
       memory: 100Mi
 
   # Enable VPA recommendations for all namespaces

grafana/helm-values.yaml

@@ -39,13 +39,12 @@ podSecurityContext:
   fsGroup: 472
   fsGroupChangePolicy: "Always"
 
-# Resource settings (VPA lowerBound/upperBound)
+# Resource settings (no CPU limit for stability)
 resources:
   requests:
     cpu: 22m
     memory: 144Mi
   limits:
-    cpu: 24m
     memory: 242Mi
 
 service:

kube-state-metrics/helm-values.yaml

@@ -7,13 +7,12 @@ fullnameOverride: kube-state-metrics
 # Note: kube-state-metrics는 stateless이지만, 여러 replica는 동일한 메트릭을 중복 생성하므로
 # 단일 replica로 실행하는 것이 권장됩니다.
 
-# Resource settings (VPA lowerBound/upperBound)
+# Resource settings (no CPU limit for stability)
 resources:
   requests:
     cpu: 15m
     memory: 100Mi
   limits:
-    cpu: 15m
     memory: 100Mi
 
 service:

loki/helm-values.yaml

@@ -60,13 +60,12 @@ singleBinary:
       mountPath: /var/loki
   # Medium priority for observability
   priorityClassName: medium-priority
-  # Resource settings (VPA lowerBound/target)
+  # Resource settings (no CPU limit for stability)
   resources:
     requests:
       cpu: 10m
       memory: 225Mi
     limits:
-      cpu: 69m
       memory: 323Mi
 
 # Disable components not needed in single binary mode

node-exporter/helm-values.yaml

@@ -6,13 +6,12 @@ fullnameOverride: node-exporter
 hostNetwork: true
 hostPID: true
 
-# Resource settings (VPA lowerBound/upperBound)
+# Resource settings (no CPU limit for stability)
 resources:
   requests:
     cpu: 15m
     memory: 100Mi
   limits:
-    cpu: 15m
     memory: 100Mi
 
 service:

opentelemetry/helm-values.yaml

@@ -28,14 +28,13 @@ image:
 mode: daemonset
 
 # =============================================================================
-# Resource Limits (VPA lowerBound/upperBound, mem limit capped at 1024Mi)
+# Resource Limits (no CPU limit for stability, mem limit capped at 1024Mi)
 # =============================================================================
 resources:
   requests:
     cpu: 34m
     memory: 142Mi
   limits:
-    cpu: 410m
     memory: 1024Mi
 
 # =============================================================================

prometheus/helm-values.yaml

@@ -14,13 +14,12 @@ prometheusOperator:
   enabled: true
   # CRD 생성 비활성화
   createCustomResource: false
-  # Resource settings (VPA lowerBound/upperBound)
+  # Resource settings (no CPU limit for stability)
   resources:
     requests:
       cpu: 15m
       memory: 100Mi
     limits:
-      cpu: 15m
       memory: 100Mi
 
 # Kubelet ServiceMonitor with cluster label

tempo/helm-values.yaml

@@ -17,13 +17,12 @@ replicas: 1
 # Tempo Configuration
 # =============================================================================
 tempo:
-  # Resource settings (VPA lowerBound/target)
+  # Resource settings (no CPU limit for stability)
   resources:
     requests:
       cpu: 15m
       memory: 100Mi
     limits:
-      cpu: 15m
       memory: 109Mi
   # Receivers - protocols Tempo accepts
   receivers:

thanos/helm-values.yaml

@@ -46,13 +46,12 @@ query:
     - --query.replica-label=prometheus_replica
     - --query.auto-downsampling
 
-  # Resource settings (VPA lowerBound/target)
+  # Resource settings (no CPU limit for stability)
   resources:
     requests:
       cpu: 15m
       memory: 100Mi
     limits:
-      cpu: 15m
       memory: 126Mi
 
 # =============================================================================

vpa/helm-values.yaml

@@ -6,13 +6,12 @@ recommender:
   enabled: true
   replicaCount: 1
 
-  # Resource settings (VPA lowerBound/upperBound)
+  # Resource settings (no CPU limit for stability)
   resources:
     requests:
       cpu: 15m
       memory: 100Mi
     limits:
-      cpu: 15m
       memory: 100Mi
 
 # Updater - applies recommended resource requests to pods