ansible/group_vars/prom/prometheus.yml

---
prometheus__alertmanager_targets:
  - docker-ovh.adm.auro.re:9093

prometheus__tsdb_retention_time: 90d

prometheus__scraping:
  node:
    targets: "{{ groups.vm_network + groups.pve_network }}"
    address:
      port: 9100
  prometheus:
    targets: "{{ groups.prom }}"
    address:
      port: 9090
  kresd:
    targets: "{{ groups.dns }}"
    address:
      port: 8453
  bird:
    targets: "{{ groups.router }}"
    address:
      port: 9324
  quanta:
    targets: "{{ groups.quanta }}"
    address: 127.0.0.1:9116
    path: /snmp
    params:
      module:
        - quanta
  snmp:
    targets: "{{ groups.prom }}"
    address:
      port: 9116

prometheus__alert_rules_prometheus:
  - alert: PrometheusTsdbCompactionFailed
    expr:
      increase(prometheus_tsdb_compactions_failed_total[1m]) > 0
    for: 0m
    labels:
      severity: critical

prometheus__alert_rules_common:
  - alert: CollectorDown
    expr: >
      up == 0
    for: 3m
    labels:
      severity: critical
    annotations:
      Job: !unsafe "{{ $labels.job }}"

prometheus__alert_rules_node:
  - alert: OutOfMemory
    expr:
      100 * (
        node_memory_MemFree_bytes
        + node_memory_Cached_bytes
        + node_memory_Buffers_bytes
      ) / node_memory_MemTotal_bytes < 10
    for: 5m
    labels:
      severity: warning
    annotations:
      FreeMemory: !unsafe '{{ printf "%.0f" $value }} %'
  - alert: HostSwapIsFillingUp
    expr:
      100 * (
        1 - (
          node_memory_SwapFree_bytes
          / node_memory_SwapTotal_bytes
        )
      ) >= 50
    for: 3m
    labels:
      severity: critical
    annotations:
      UsedSwap: !unsafe '{{ printf "%.0f" $value }} %'
  - alert: HostPhysicalComponentTooHot
    expr:
      node_hwmon_temp_celsius > 79
    for: 3m
    labels:
      severity: critical
    annotations:
      Temperature: !unsafe '{{ printf "%.0f" $value }} °C'
      Chip: !unsafe "{{ $labels.chip }}"
      Sensor: !unsafe "{{ $labels.sensor }}"
  - alert: HostNodeOvertemperatureAlarm
    expr:
      node_hwmon_temp_crit_alarm_celsius == 1
    for: 0m
    labels:
      severity: critical
    annotations:
      Chip: !unsafe "{{ $labels.chip }}"
      Sensor: !unsafe "{{ $labels.sensor }}"
  - alert: HostRaidArrayGotInactive
    expr:
      node_md_state{state="inactive"} > 0
    for: 0m
    labels:
      severity: critical
    annotations:
      Device: !unsafe "{{ $labels.device }}"
  - alert: HostRaidDiskFailure
    expr:
      node_md_disks{state="failed"} > 0
    for: 0m
    labels:
      severity: critical
    annotations:
      severity: !unsafe "{{ $labels.md_device }}"
  - alert: HostOomKillDetected
    expr:
      increase(node_vmstat_oom_kill[1m]) > 0
    for: 0m
    labels:
      severity: warning
    annotations:
      PID: !unsafe "{{ $value }}"
  - alert: HostEdacCorrectableErrorsDetected
    expr:
      increase(node_edac_correctable_errors_total[1m]) > 0
    for: 0m
    labels:
      severity: warning
    annotations:
      CorrectedErrors: !unsafe "{{ $value }}"
  - alert: HostEdacUncorrectableErrorsDetected
    expr:
      increase(node_edac_uncorrectable_errors_total[1m]) > 0
    for: 0m
    labels:
      severity: warning
    annotations:
      DetectedErrors: !unsafe "{{ $value }}"
  - alert: OutOfDiskSpace
    expr:
      (
        100 * node_filesystem_free_bytes
        / node_filesystem_size_bytes < 10
      )
      and on (instance, device, mountpoint) (
        node_filesystem_readonly
      ) == 0
    for: 5m
    labels:
      severity: critical
    annotations:
      Mountpoint: !unsafe '{{ $labels.mountpoint }}'
      FreeSpace: !unsafe '{{ printf "%.0f" $value }} %'
  - alert: HostConntrackLimit
    expr:
      100 * (
        node_nf_conntrack_entries
        / node_nf_conntrack_entries_limit
      ) > 80
    for: 5m
    labels:
      severity: warning
    annotations:
      Filled: !unsafe '{{ printf "%.0f" $value }} %'
  - alert: HostClockSkew
    expr:
      (
        node_timex_offset_seconds > 0.05
        and deriv(node_timex_offset_seconds[5m]) >= 0
      ) or (
        node_timex_offset_seconds < -0.05
        and deriv(node_timex_offset_seconds[5m]) <= 0
      )
    for: 2m
    labels:
      severity: warning
  - alert: HostClockNotSynchronising
    expr:
      min_over_time(node_timex_sync_status[1m]) == 0
      and node_timex_maxerror_seconds >= 16
    for: 2m
    labels:
      severity: warning
  - alert: HostRequiresReboot
    expr:
      node_reboot_required > 0
    for: 5m
    labels:
      severity: warning
  - alert: OutOfInodes
    expr:
      100 * node_filesystem_files_free
      / node_filesystem_files < 10
    for: 3m
    labels:
      severity: warning
    annotations:
      Mountpoint: !unsafe "{{ $labels.mountpoint }}"
      FreeInodes: !unsafe '{{ printf "%.0f" $value }} %'
  - alert: CpuUsage
    expr:
      100 * (
        1 - avg by (instance) (
          irate(node_cpu_seconds_total{mode="idle"}[5m])
        )
      ) > 75
    for: 10m
    labels:
      severity: warning
    annotations:
      Usage: !unsafe '{{ printf "%.0f" $value }} %'
  - alert: SystemdServiceFailed
    expr:
      node_systemd_unit_state{state="failed"} == 1
    for: 10m
    labels:
      severity: warning
    annotations:
      Service: !unsafe "{{ $labels.name }}"
  - alert: LoadUsage
    expr:
      node_load1 > 5
    for: 2m
    labels:
      severity: warning
    annotations:
      Load1: !unsafe '{{ printf "%.0f" $value }}'
  - alert: UnhealthyDisk
    expr:
      smartmon_device_smart_healthy < 1
    for: 10m
    labels:
      severity: critical
    annotations:
      Disk: !unsafe "{{ $labels.disk }}"
  - alert: HostCpuStealNoisyNeighbor
    expr:
      100 * avg by (instance) (
        rate(node_cpu_seconds_total{mode="steal"}[5m])
      ) > 10
    for: 5m
    labels:
      severity: warning
    annotations:
      Disk: !unsafe "{{ $labels.disk }}"

prometheus__alert_rules_keepalived:
  - alert: KeepalivedVrrpFault
    expr:
      keepalived_vrrp_state{state="fault"} > 0
    for: 0m
    labels:
      severity: critical
    annotations:
      Instance: !unsafe "{{ $labels.instance }}"
  - alert: KeepalivedMasterChange
    expr:
      changes(
        keepalived_vrrp_state{
          keepalived_vvrp_state="master"
        }[1m]
      ) > 1
    for: 0m
    labels:
      severity: warning
    annotations:
      Instance: !unsafe "{{ $labels.instance }}"

prometheus__alert_rules_bird:
  - record: bird:protocol_up:bgp
    expr:
      label_replace(
        bird_protocol_up{proto="BGP",}
        unless bird_protocol_up{
          proto="BGP",
          name=~"^(viarezo|isp[12]|rezel)[46]$"
        },
        "group", "$1",
        "instance", "^([^0-9\\.]+)-[0-9]+.*"
      )
  # Sessions qui ne sont volontairement pas redondées
  # au sein d'un groupe
  - record: bird:protocol_up:bgp:non_redundant
    expr:
      bird:protocol_up:bgp{
        group="edge",
        name=~"^(oti|crans|legacy|edge)[46]$"
      }
  # Sessions qui le sont
  - record: bird:protocol_up:bgp:redundant
    expr:
      bird:protocol_up:bgp
      unless
      bird:protocol_up:bgp:non_redundant
  - alert: BirdBGPRedundancyDegraded
    expr:
      (
        count by (group, name) (
          bird:protocol_up:bgp:redundant{state="Established"}
        ) or (
          count by (group, name) (
            bird:protocol_up:bgp:redundant{state!="Established"}
          ) * 0
        )
      ) < 2
    for: 0m
    labels:
      severity: warning
    annotations:
      Session: !unsafe "{{ $labels.name }}"
      Count: !unsafe "{{ $value }}"
      Group: !unsafe "{{ $labels.group }}"
  - alert: BirdBGPDown
    expr:
      (
        count by (group, name) (
          bird:protocol_up:bgp{state="Established"}
        ) or (
          count by (group, name) (
            bird:protocol_up:bgp{state!="Established"}
          ) * 0
        )
      ) == 0
    for: 0m
    labels:
      severity: critical
    annotations:
      Session: !unsafe "{{ $labels.name }}"
      Group: !unsafe "{{ $labels.group }}"
  - alert: BirdBGPNoExportedPrefixRedundant
    expr:
      bird_protocol_prefix_export_count{
        export_filter!="REJECT",
      } * on (instance, name) (
        bird:protocol_up:bgp:redundant{state="Established"}
      ) == 0
    for: 0m
    labels:
      severity: critical
    annotations:
      Session: !unsafe "{{ $labels.name }}"
  - alert: BirdBGPNoImportedPrefixRedundant
    expr:
      bird_protocol_prefix_import_count{
        import_filter!="REJECT",
      } * on (instance, name) (
        bird:protocol_up:bgp:redundant{state="Established"}
      ) == 0
    for: 0m
    labels:
      severity: critical
    annotations:
  - alert: BirdBGPNoExportedPrefixNonRedundant
    expr:
      sum by (group) (
        bird_protocol_prefix_export_count{
          export_filter!="REJECT",
        } * on (instance, name) group_left (group) (
          bird:protocol_up:bgp:non_redundant{state="Established"}
        )
      ) == 0
    for: 0m
    labels:
      severity: critical
    annotations:
      Session: !unsafe "{{ $labels.name }}"
  - alert: BirdBGPNoImportedPrefixNonRedundant
    expr:
      sum by (group) (
        bird_protocol_prefix_import_count{
          import_filter!="REJECT",
        } * on (instance, name) group_left (group) (
          bird:protocol_up:bgp:non_redundant{state="Established"}
        )
      ) == 0
    for: 0m
    labels:
      severity: critical
    annotations:
      Session: !unsafe "{{ $labels.name }}"
  - alert: BirdOSPFNeighboursChange
    expr:
      changes(bird_ospf_neighbor_count[5m]) > 0
      or changes(bird_ospfv3_neighbor_count[5m]) > 0
    for: 0m
    labels:
      severity: warning
  - alert: BirdOSPFDown
    expr:
      bird_ospf_running == 0
    for: 0m
    labels:
      severity: critical
    annotations:
      Instance: !unsafe "{{ $labels.name }}"

prometheus__alert_rules_quanta:
  - alert: QuantaQueueOverflow
    expr:
      snAgGblQueueOverflow == 1
    for: 0m
    labels:
      severity: critical
  - alert: QuantaCpuUsage
    expr:
      snAgGblCpuUtil1MinAvg > 50
    for: 5m
    labels:
      severity: warning
    annotations:
      Usage: !unsafe "{{ $value }} %"
  - alert: QuantaCpuUsage
    expr:
      snAgGblCpuUtil1MinAvg > 80
    for: 5m
    labels:
      severity: critical
    annotations:
      Usage: !unsafe "{{ $value }} %"
  - alert: QuantaMemoryUsage
    expr:
      100 * (1 - (snAgGblDynMemFree / snAgGblDynMemTotal)) > 50
    for: 5m
    labels:
      severity: warning
    annotations:
      UsedMemory: !unsafe "{{ $value }} %"
  - alert: QuantaMemoryUsage
    expr:
      100 * (1 - (snAgGblDynMemFree / snAgGblDynMemTotal)) > 80
    for: 5m
    labels:
      severity: alert
    annotations:
      UsedMemory: !unsafe "{{ $value }} %"
  - alert: QuantaFanHealth
    expr:
      snChasFanOperStatus{snChasFanOperStatus="normal"} == 0
    for: 0m
    labels:
      severity: critical
    annotations:
      Description: !unsafe "{{ $labels.shChasFanDescription }}"
      Status: !unsafe "{{ $labels.snChasFanOperStatus }}"
  - alert: QuantaTemp
    expr:
      0.5 * snAgentTempValue > 45
    for: 0m
    labels:
      severity: warning
    annotations:
      Temperature: !unsafe "{{ $value }} °C"
      Description: !unsafe "{{ $labels.snAgentTempSensorDescr }}"
  - alert: QuantaTemp
    expr:
      0.5 * snAgentTempValue > 60
    for: 0m
    labels:
      severity: critical
    annotations:
      Temperature: !unsafe "{{ $value }} °C"
      Description: !unsafe "{{ $labels.snAgentTempSensorDescr }}"
  - alert: QuantaPowerRedundancyFailure
    expr:
      count by (instance) (
        snChasPwrSupplyOperStatus{snChasPwrSupplyOperStatus="normal"}
      ) < 2
    for: 0m
    labels:
      severity: warning

prometheus__alert_rules_switch:
  - alert: SwitchPromiscuousChange
    expr:
      changes(ifPromiscuousMode[5m]) > 0
    for: 0m
    labels:
      severity: warning
    annotations:
      Interface: !unsafe "{{ $labels.ifName }}
        {{ if $labels.ifAlias }}- {{ $labels.ifAlias }}{{ end }}"
  - alert: SwitchInterfaceUpChange
    expr:
      changes(ifOperStatus{ifOperStatus="up"}[5m]) > 0
    for: 0m
    labels:
      severity: warning
    annotations:
      Interface: !unsafe "{{ $labels.ifName }}
        {{ if $labels.ifAlias }}- {{ $labels.ifAlias }}{{ end }}"
  - alert: SwitchInErrors
    expr:
      100 * irate(ifInErrors[5m]) / (
        irate(ifInUcastPkts[5m])
        + irate(ifInNUcastPkts[5m])
      ) > 0.01
    for: 0m
    labels:
      severity: warning
    annotations:
      ErrorRate: !unsafe '{{ printf "%.2f" $value }} %'
      Interface: !unsafe "{{ $labels.ifName }}
        {{ if $labels.ifAlias }}- {{ $labels.ifAlias }}{{ end }}"
  - alert: SwitchOutErrors
    expr:
      100 * irate(ifOutErrors[5m]) / (
        irate(ifOutUcastPkts[5m])
        + irate(ifOutNUcastPkts[5m])
      ) > 0.01
    for: 0m
    labels:
      severity: warning
    annotations:
      ErrorRate: !unsafe '{{ printf "%.2f" $value }} %'
      Interface: !unsafe "{{ $labels.ifName }}
        {{ if $labels.ifAlias }}- {{ $labels.ifAlias }}{{ end }}"

prometheus__alert_rules:
  common: "{{ prometheus__alert_rules_common }}"
  switch: "{{ prometheus__alert_rules_switch }}"
  prometheus: "{{ prometheus__alert_rules_prometheus }}"
  node: "{{ prometheus__alert_rules_node }}"
  keepalived: "{{ prometheus__alert_rules_keepalived }}"
  quanta: "{{ prometheus__alert_rules_quanta }}"
  bird: "{{ prometheus__alert_rules_bird }}"
...