prepare for production

k8s/monitoring/templates/node-alerts.yaml

@@ -10,10 +10,17 @@ spec:
   - name: node.alerts
     rules:
     - alert: NodeHighCPU
-      expr: (1 - avg without (cpu, mode) (rate(node_cpu_seconds_total{mode="idle"}[1m]))) * 100 > 1
+      expr: |
+        (
+          (1 - avg without (cpu, mode) (rate(node_cpu_seconds_total{mode="idle", workload="media"}[1m]))) * 100 > 65
+        )
+        or
+        (
+          (1 - avg without (cpu, mode) (rate(node_cpu_seconds_total{mode="idle", workload!="media"}[1m]))) * 100 > 80
+        )
       for: 1m
       labels:
         severity: warning
       annotations:
         summary: "High CPU usage on node {{ "{{" }} $labels.instance {{ "}}" }}"
-        description: "Node {{ "{{" }} $labels.instance {{ "}}" }} has CPU usage above 1% (current value: {{ "{{" }} $value | printf \"%.2f\" {{ "}}" }}%)"
+        description: "Node {{ "{{" }} $labels.instance {{ "}}" }} has CPU usage above threshold (current value: {{ "{{" }} $value | printf \"%.2f\" {{ "}}" }}%)"

k8s/webrtc-be/values-production.yaml

@@ -6,7 +6,7 @@ grafana:
 backendUrl: "https://www.jamkazam.com"
 auth: true
 # Webrtc_be image version
-webrtc_tag: 1.0.139
+webrtc_tag: 1.0.157
 unused_tag: 1.0.0
 enableDieEndpoint: false
 mediasoup:

terraform/lke.tf

@@ -71,7 +71,7 @@ resource "linode_lke_cluster" "prd-video-cluster" {
   }
   # Coturn pool
   pool {
-    type  = "g6-standard-2"
+    type  = "g6-standard-4"
     count = 3
     labels = {
         workload = "media"