alertmanager_rules_for_Linux_Windows.yml

groups:
- name: alert-rules
  rules:
  - alert: ExporterDown
    expr: up == 0
    for: 2m
    labels:
      severity: critical
    annotations:
      description: "Node {{ $labels.job }} running on {{ $labels.instance }} has been down"

  - alert: Host Out Of DiskSpace
    expr: 100 - ((node_filesystem_avail_bytes{mountpoint="/",fstype!="tmpfs"} * 100) / node_filesystem_size_bytes{mountpoint="/",fstype!="tmpfs"}) > 85
    for: 2m
    labels:
      severity: warning
    annotations:
      description: "Disk filled\n  VALUE = {{ $value }}"


  - alert: Host Out Of Memory
    expr: node_memory_MemAvailable_bytes / node_memory_MemTotal_bytes * 100 < 15
    for: 5m
    labels:
      severity: warning
    annotations:
      description: "Memory remaining\n  VALUE = {{ $value }}"

  - alert: HostHighCpuLoad
    expr: 100 - (avg by(instance) (rate(node_cpu_seconds_total{mode="idle"}[2m])) * 100) > 85
    for: 5m
    labels:
      severity: warning
    annotations:
      description: "CPU load\n  VALUE = {{ $value }}"

  - alert: Windows Server Collector Error
    expr: wmi_exporter_collector_success == 0
    for: 0m
    labels:
      severity: critical
    annotations:
      description: "Collector {{ $labels.collector }} was not successful\n  VALUE = {{ $value }}\n  LABELS = {{ $labels }}"

  - alert: Windows Server Service Status
    expr: wmi_service_status{status="ok"} != 1
    for: 1m
    labels:
      severity: critical
    annotations:
      description: "Windows Service state is not OK\n  VALUE = {{ $value }}"

  - alert: Windows Server CPU Usage
    expr: 100 - (avg by (instance) (rate(wmi_cpu_time_total{mode="idle"}[2m])) * 100) > 85
    for: 5m
    labels:
      severity: warning
    annotations:
      description: "CPU Usage is more than\n  VALUE = {{ $value }}"

  - alert: Windows Server Memory Usage
    expr: 100 - ((wmi_os_physical_memory_free_bytes / wmi_cs_physical_memory_bytes) * 100) > 85
    for: 5m
    labels:
      severity: warning
    annotations:
      description: "Memory usage is more than\n  VALUE = {{ $value }}"
  
  - alert: Windows Server DiskSpace Usage
    expr: 100.0 - 100 * ((wmi_logical_disk_free_bytes / 1024 / 1024 ) / (wmi_logical_disk_size_bytes / 1024 / 1024)) > 85
    for: 5m
    labels:
      severity: critical
    annotations:
      description: "Disk usage is more than\n  VALUE = {{ $value }}"