Inital Commit - Prometheus + Grafana

.gitignore

@@ -0,0 +1,2 @@
+*
+!docker-compose.yml

config/alertmanager.yml

@@ -0,0 +1,20 @@
+global:
+  resolve_timeout: 5m
+
+route:
+  group_by: ['Alertname']
+  group_wait: 10s
+  group_interval: 10s
+  repeat_interval: 24h
+  receiver: 'email'
+receivers:
+- name: 'email'
+  email_configs:
+  - to: 'matanhorovitz@protonmail.com'
+    from: 'pukekoxyz@yahoo.com'
+    smarthost: smtp.mail.yahoo.com:587
+    auth_username: 'pukekoxyz@yahoo.com'
+    auth_identity: 'pukekoxyz@yahoo.com'
+    auth_password: 'pvefngmuhcxunzqs'
+    require_tls: yes
+    send_resolved: true

config/alerts.yml

@@ -0,0 +1,14 @@
+groups:
+  - name: Uptime
+    rules:
+      - alert: InstanceDown 
+        expr: up{job="services"} < 1 
+        for: 5m 
+  - name: Usage
+    rules:
+      - alert: HighRootFSDiskUsage
+        expr: 100 - ((node_filesystem_avail_bytes{mountpoint="/",fstype!="rootfs"} * 100) / node_filesystem_size_bytes{mountpoint="/",fstype!="rootfs"}) > 80
+        for: 1m
+      - alert: HighRedVolDiskUsage
+        expr: 100 - ((node_filesystem_avail_bytes{mountpoint="/Red-Vol",fstype!="rootfs"} * 100) / node_filesystem_size_bytes{mountpoint="/Red-Vol",fstype!="rootfs"}) > 70
+        for: 1m

config/prometheus.yml

@@ -0,0 +1,23 @@
+global:
+  scrape_interval:     15s
+  evaluation_interval: 15s
+
+alerting:
+  alertmanagers:
+  - static_configs:
+    - targets: ['alertmanager:9093']
+rule_files:
+  - alerts.yml
+
+scrape_configs:
+  - job_name: prometheus
+    static_configs:
+      - targets: ['prometheus:9090']
+  - job_name: takahe
+    static_configs:
+      - targets: ['192.168.0.66:9100']
+  - job_name: cadvisor
+    scrape_interval: 5s
+    static_configs:
+    - targets:
+      - cadvisor:8080

docker-compose.yml

@@ -0,0 +1,89 @@
+version: '3.2'
+services:
+  prometheus:
+    image: prom/prometheus:latest
+    container_name: prometheus
+    ports:
+    - 9090:9090
+    command:
+    - --config.file=/etc/prometheus/prometheus.yml
+    volumes:
+    - ./config/prometheus.yml:/etc/prometheus/prometheus.yml:ro
+    - ./config/alerts.yml:/etc/prometheus/alerts.yml
+    depends_on:
+    - cadvisor
+    restart: unless-stopped
+    networks:
+      - network
+      - internal
+    labels:
+      - "traefik.enable=true"
+      - "traefik.docker.network=prometheus_network"
+      - "traefik.http.routers.prometheus.entrypoints=websecure"
+      - "traefik.http.routers.prometheus.rule=Host(`monitor.pukeko.xyz`)"
+      - "traefik.http.routers.prometheus.tls.certresolver=pukekoresolver"
+      - "traefik.http.routers.prometheus.middlewares=authelia@docker"
+
+  alertmanager:
+    container_name: alertmanager
+    image: prom/alertmanager
+    volumes:
+    - ./config/alerts.yml:/etc/prometheus/alerts.yml
+    - ./config/alertmanager.yml:/etc/prometheus/alertmanager.yml
+    ports:
+    - '9093:9093'
+    restart: unless-stopped
+    networks:
+      - internal
+  cadvisor:
+    image: gcr.io/cadvisor/cadvisor:latest
+    container_name: cadvisor
+    ports:
+    - 1010:8080
+    volumes:
+    - /:/rootfs:ro
+    - /var/run:/var/run:rw
+    - /sys:/sys:ro
+    - /var/lib/docker/:/var/lib/docker:ro
+    depends_on:
+    - redis
+    restart: unless-stopped
+    networks:
+      - internal
+  redis:
+    image: redis:alpine
+    container_name: redis
+    ports:
+    - 6379:6379
+    restart: unless-stopped
+    networks:
+      - internal
+  grafana:
+    container_name: grafana
+    image: grafana/grafana
+    depends_on:
+    - prometheus
+    ports:
+    - '1000:3000'
+    volumes:
+    - './grafana/data:/var/lib/grafana'
+    - './grafana/provisioning/:/etc/grafana/provisioning/'
+    - './grafana/config/:/etc/grafana/'
+    restart: unless-stopped
+    user: '1000'
+    networks:
+      - network
+      - internal
+    labels:
+      - "traefik.enable=true"
+      - "traefik.docker.network=prometheus_network"
+      - "traefik.http.routers.grafana.entrypoints=websecure"
+      - "traefik.http.services.grafana.loadbalancer.server.port=3000"
+      - "traefik.http.routers.grafana.rule=Host(`flight.pukeko.xyz`)"
+      - "traefik.http.routers.grafana.tls.certresolver=pukekoresolver"
+      - "traefik.http.routers.grafana.middlewares=authelia@docker"
+networks:
+  network:
+    driver: bridge
+  internal:
+    driver: bridge