From 8af179f7a63265d3b3f5f3b8e3af542b0537a022 Mon Sep 17 00:00:00 2001
From: renolation <vodanh.2901@gmail.com>
Date: Fri, 19 Jun 2026 17:33:56 +0700
Subject: [PATCH] add monitor stack

---
 envs/dev/apps/monitoring.yaml   | 29 ++++++++++
 envs/dev/values/monitoring.yaml | 93 +++++++++++++++++++++++++++++++++
 2 files changed, 122 insertions(+)
 create mode 100644 envs/dev/apps/monitoring.yaml
 create mode 100644 envs/dev/values/monitoring.yaml

diff --git a/envs/dev/apps/monitoring.yaml b/envs/dev/apps/monitoring.yaml
new file mode 100644
index 0000000..4e7c42e
--- /dev/null
+++ b/envs/dev/apps/monitoring.yaml
@@ -0,0 +1,29 @@
+apiVersion: argoproj.io/v1alpha1
+kind: Application
+metadata:
+  name: monitoring-dev
+  namespace: argocd
+spec:
+  project: default
+  sources:
+    # Chart từ Helm repo prometheus-community (không cần đưa vào Gitea)
+    - repoURL: https://prometheus-community.github.io/helm-charts
+      chart: kube-prometheus-stack
+      targetRevision: 66.2.1
+      helm:
+        valueFiles:
+          - $values/envs/dev/values/monitoring.yaml
+    # Repo Gitea chỉ để lấy values
+    - repoURL: https://git.renolation.com/renolation/argo-k8s.git
+      targetRevision: main
+      ref: values
+  destination:
+    server: https://kubernetes.default.svc
+    namespace: monitoring
+  syncPolicy:
+    automated:
+      prune: true
+      selfHeal: true
+    syncOptions:
+      - CreateNamespace=true
+      - ServerSideApply=true   # CRD kube-prometheus-stack rất lớn, cần server-side
\ No newline at end of file
diff --git a/envs/dev/values/monitoring.yaml b/envs/dev/values/monitoring.yaml
new file mode 100644
index 0000000..aa2ca5e
--- /dev/null
+++ b/envs/dev/values/monitoring.yaml
@@ -0,0 +1,93 @@
+# kube-prometheus-stack — DEV values
+# Gọn cho dev: dung lượng nhỏ, expose Grafana qua Traefik ingress.
+
+# ---------------------------------------------------------------------------
+# Grafana
+# ---------------------------------------------------------------------------
+grafana:
+  enabled: true
+  # Mật khẩu admin ban đầu (đổi sau khi vào). Dev nên để đây cho nhanh;
+  # production thì chuyển sang Sealed Secret.
+  adminPassword: "Dbiz1234"
+  ingress:
+    enabled: true
+    ingressClassName: traefik
+    hosts:
+      - grafana.dev.dbiz.local
+    paths:
+      - /
+  persistence:
+    enabled: true
+    size: 2Gi
+  # Dashboard K8s mặc định đã có sẵn trong chart
+
+# ---------------------------------------------------------------------------
+# Prometheus
+# ---------------------------------------------------------------------------
+prometheus:
+  ingress:
+    enabled: true
+    ingressClassName: traefik
+    hosts:
+      - prometheus.dev.dbiz.local
+    paths:
+      - /
+  prometheusSpec:
+    retention: 7d
+    # Cho phép Prometheus tự phát hiện ServiceMonitor/PodMonitor ở mọi namespace
+    serviceMonitorSelectorNilUsesHelmValues: false
+    podMonitorSelectorNilUsesHelmValues: false
+    ruleSelectorNilUsesHelmValues: false
+    storageSpec:
+      volumeClaimTemplate:
+        spec:
+          accessModes: ["ReadWriteOnce"]
+          resources:
+            requests:
+              storage: 10Gi
+    resources:
+      requests:
+        cpu: 200m
+        memory: 512Mi
+      limits:
+        memory: 1Gi
+
+# ---------------------------------------------------------------------------
+# Alertmanager
+# ---------------------------------------------------------------------------
+alertmanager:
+  enabled: true
+  ingress:
+    enabled: true
+    ingressClassName: traefik
+    hosts:
+      - alertmanager.dev.dbiz.local
+    paths:
+      - /
+  alertmanagerSpec:
+    storage:
+      volumeClaimTemplate:
+        spec:
+          accessModes: ["ReadWriteOnce"]
+          resources:
+            requests:
+              storage: 2Gi
+
+# ---------------------------------------------------------------------------
+# Exporters — bật để có metrics node + kube-state
+# ---------------------------------------------------------------------------
+nodeExporter:
+  enabled: true
+kubeStateMetrics:
+  enabled: true
+
+# Một số control-plane component khó scrape trên cluster kubeadm tự dựng;
+# tắt scrape etcd/scheduler/controller nếu chưa expose metrics để khỏi báo đỏ.
+kubeEtcd:
+  enabled: false
+kubeControllerManager:
+  enabled: false
+kubeScheduler:
+  enabled: false
+kubeProxy:
+  enabled: false
\ No newline at end of file