feat: add health checks

docker-compose.yaml

@@ -1,7 +1,3 @@
-# https://docs.docker.com/reference/compose-file/version-and-name
-# https://docs.docker.com/reference/compose-file/services
-# https://docs.docker.com/reference/compose-file/volumes
-
 name: monitoring
 
 services:
@@ -15,6 +11,13 @@ services:
     volumes:
       - ./loki.yaml:/etc/loki/loki.yaml
       - loki-data:/loki
+    healthcheck:
+      test: ["CMD", "curl", "http://localhost:3100/ready"]
+      interval: 30s
+      timeout: 5s
+      retries: 3
+      start_period: 5s
+
   node-exporter:
     image: prom/node-exporter:v1.8.2
     container_name: node-exporter
@@ -28,26 +31,43 @@ services:
       - /proc:/host/proc:ro
       - /sys:/host/sys:ro
       - /:/host:ro
+    healthcheck:
+      test: ["CMD", "curl", "http://localhost:9100/metrics"]
+      interval: 30s
+      timeout: 5s
+      retries: 3
+      start_period: 5s
+
   prometheus:
     image: prom/prometheus:v3.0.0
     container_name: prometheus
     restart: unless-stopped
     env_file: .env
     depends_on:
-      - node-exporter
+      node-exporter:
+        condition: service_healthy
     expose:
       - 9090
     volumes:
       - ./prometheus.yaml:/etc/prometheus/prometheus.yml
       - prometheus-data:/prometheus
+    healthcheck:
+      test: ["CMD", "curl", "http://localhost:9090/-/ready"]
+      interval: 30s
+      timeout: 5s
+      retries: 3
+      start_period: 5s
+
   grafana:
     image: grafana/grafana-oss:11.3.1
     container_name: grafana
     restart: unless-stopped
     env_file: .env
     depends_on:
-      - loki
-      - prometheus
+      loki:
+        condition: service_healthy
+      prometheus:
+        condition: service_healthy
     ports:
       - ${GRAFANA_PORT}:3000
     volumes:
@@ -56,6 +76,12 @@ services:
       - ./grafana-metrics-dashboard.json:/var/lib/grafana/dashboards/metrics.json
       - ./grafana-logs-dashboard.json:/var/lib/grafana/dashboards/logs.json
       - grafana-data:/var/lib/grafana
+    healthcheck:
+      test: ["CMD", "curl", "http://localhost:3000/api/health"]
+      interval: 30s
+      timeout: 5s
+      retries: 3
+      start_period: 5s
 
 volumes:
   loki-data:

grafana-datasources.yaml

@@ -8,7 +8,7 @@ datasources:
     access: proxy
 
     url: http://prometheus:9090
-  - name: Loki
+  - name: DS_LOKI
     type: loki
     access: proxy