Merge pull request 'Cleanup Prometheus alerts' (#45) from prometheus_alerts into master
All checks were successful
continuous-integration/drone/push Build is passing

Reviewed-on: Aurore/ansible#45
This commit is contained in:
jeltz 2021-04-01 19:24:05 +02:00
commit 2388bfdc3d
11 changed files with 185 additions and 170 deletions

6
hosts
View file

@ -29,7 +29,6 @@ stream.adm.auro.re
re2o-server.adm.auro.re
re2o-ldap.adm.auro.re
re2o-db.adm.auro.re
backup.adm.auro.re
mail.adm.auro.re
wikijs.adm.auro.re
prometheus-aurore.adm.auro.re
@ -39,7 +38,6 @@ log.adm.auro.re
bdd.adm.auro.re
bdd-ovh.adm.auro.re
litl.adm.auro.re
services-bdd-local.adm.auro.re
log.adm.auro.re
[aurore_testing_vm]
@ -53,7 +51,6 @@ horus.adm.auro.re
[ovh_container]
synapse.adm.auro.re
#services-bdd.adm.auro.re
phabricator.adm.auro.re
wiki.adm.auro.re
www.adm.auro.re
@ -63,7 +60,6 @@ matrix-services.adm.auro.re
[ovh_vm]
serge.adm.auro.re
passbolt.adm.auro.re
vpn-ovh.adm.auro.re
docker-ovh.adm.auro.re
switchs-manager.adm.auro.re
ldap-replica-ovh.adm.auro.re
@ -252,7 +248,6 @@ perceval.adm.auro.re
[edc_pve]
chapalux.adm.auro.re
escalope.adm.auro.re
[edc_vm]
routeur-edc.adm.auro.re
@ -516,5 +511,4 @@ proxy.adm.auro.re
[bdd]
bdd.adm.auro.re
bdd-ovh.adm.auro.re
services-bdd-local.adm.auro.re
re2o-db.adm.auro.re

View file

@ -8,3 +8,4 @@
service:
name: prometheus-snmp-exporter
state: restarted
...

View file

@ -11,20 +11,16 @@
- name: Configure Prometheus
template:
src: prometheus/prometheus.yml.j2
dest: /etc/prometheus/prometheus.yml
mode: 0644
notify: Restart Prometheus
- name: Configure Prometheus alert rules
template:
src: "prometheus/{{ item }}.j2"
src: "{{ item }}.j2"
dest: "/etc/prometheus/{{ item }}"
mode: 0644
notify: Restart Prometheus
owner: prometheus
group: prometheus
mode: u=r,g=r,o=
loop:
- prometheus.yml
- alert.rules.yml
- django.rules.yml
notify: Restart Prometheus
- name: Make Prometheus snmp-exporter listen on localhost only
lineinfile:
@ -36,10 +32,11 @@
# This file store SNMP OIDs
- name: Configure Prometheus snmp-exporter
template:
src: "prometheus/snmp.yml.j2"
dest: "/etc/prometheus/snmp.yml"
mode: 0600
src: snmp.yml.j2
dest: /etc/prometheus/snmp.yml
owner: prometheus
group: prometheus
mode: u=r,g=r,o=
notify: Restart prometheus-snmp-exporter
# We don't need to restart Prometheus when updating nodes

View file

@ -0,0 +1,160 @@
---
{{ ansible_managed | comment }}
{% macro raw(string) -%}
{{ "{{" }} {{ string }} {{ "}}" }}
{%- endmacro %}
groups:
- name: alert.rules
rules:
- alert: InstanceDown
expr: up == 0
for: 3m
labels:
severity: critical
annotations:
summary: >-
Invisible depuis plus de 3 minutes
- alert: OutOfMemory
expr: >-
(
node_memory_MemFree_bytes
+ node_memory_Cached_bytes
+ node_memory_Buffers_bytes
) / node_memory_MemTotal_bytes * 100 < 10
for: 5m
labels:
severity: warning
annotations:
summary: >-
{{ raw('$value | printf "%.1f"') }}% de mémoire
libre
- alert: OutOfDiskSpace
expr: >-
node_filesystem_free_bytes / node_filesystem_size_bytes * 100 < 10
for: 5m
labels:
severity: warning
annotations:
summary: >-
{{ raw('$value | printf "%.1f"') }}% d'espace libre pour
{{ raw('$labels.mountpoint') }}
- alert: OutOfInodes
expr: node_filesystem_files_free / node_filesystem_files * 100 < 10
for: 5m
labels:
severity: warning
annotations:
summary: >-
{{ raw('$value | printf "%.1f"') }}% d'inodes
restants pour {{ raw('$labels.mountpoint') }}
- alert: CpuUsage
expr: >-
(
100 - avg by (instance) (
irate(node_cpu_seconds_total{mode="idle"}[5m])
) * 100
) > 75
for: 10m
labels:
severity: warning
annotations:
summary: >-
CPU à {{ raw('$value | printf "%.1f"') }}%
- alert: SystemdServiceFailed
expr: node_systemd_unit_state{state="failed"} == 1
for: 10m
labels:
severity: warning
annotations:
summary: >-
{{ raw('$labels.name') }} a échoué
- alert: LoadUsage
expr: node_load1 > 5
for: 2m
labels:
severity: warning
annotations:
summary: >
Charge à {{ raw('$value') }}
- alert: UpsOutputSourceChanged
expr: upsOutputSource != 3
for: 1m
labels:
severity: critical
annotations:
summary: >-
Source d'alimentation changée
- alert: UpsBatteryStatus
expr: upsBatteryStatus == 3
for: 2m
labels:
severity: warning
annotations:
summary: >-
État de la batterie faible
- alert: UpsBatteryStatus
expr: upsBatteryStatus == 4
for: 10m
labels:
severity: critical
annotations:
summary: >-
État de la batterie critique
- alert: UpsHighLoad
expr: upsOutputPercentLoad > 70
for: 5m
labels:
severity: critical
annotations:
summary: >-
Charge de {{ raw('$value | printf "%.1f"') }}%
- alert: UpsWrongInputVoltage
expr: (upsInputVoltage < 210) or (upsInputVoltage > 250)
for: 10m
labels:
severity: warning
annotations:
summary: >-
Tension d'entrée de {{ raw('$value') }}V
- alert: UpsWrongOutputVoltage
expr: (upsOutputVoltage < 220) or (upsOutputVoltage > 240)
for: 10m
labels:
severity: warning
annotations:
summary: >-
Tension de sortie de {{ raw('$value') }}V
- alert: UpsTimeRemaining
expr: upsEstimatedMinutesRemaining < 8
for: 1m
labels:
severity: warning
annotations:
summary: >-
Autonomie restante de {{ raw('$value') }} min
- alert: UpsTimeRemaining
expr: upsEstimatedMinutesRemaining < 5
for: 1m
labels:
severity: critical
annotations:
summary: >-
Autonomie restante de {{ raw('$value') }} min
...

View file

@ -1,4 +1,5 @@
# {{ ansible_managed }}
---
{{ ansible_managed | comment }}
global:
# scrape_interval is set to the global default (60s)
@ -100,3 +101,4 @@ scrape_configs:
file_sd_configs:
- files:
- '/etc/prometheus/targets_docker.json'
...

View file

@ -1,138 +0,0 @@
# {{ ansible_managed }}
{# As this is also Jinja2 it will conflict without a raw block #}
{# Depending of Prometheus Node exporter version, rules can change depending of version #}
{% raw %}
groups:
- name: alert.rules
rules:
# Alert for any instance that is unreachable for >3 minutes.
- alert: InstanceDown
expr: up == 0
for: 3m
labels:
severity: critical
annotations:
summary: "{{ $labels.instance }} est invisible depuis plus de 3 minutes !"
# Alert for out of memory
- alert: OutOfMemory
expr: (node_memory_MemFree_bytes + node_memory_Cached_bytes + node_memory_Buffers_bytes) / node_memory_MemTotal_bytes * 100 < 10
for: 5m
labels:
severity: warning
annotations:
summary: "Mémoire libre de {{ $labels.instance }} à {{ humanize $value }}%."
# Alert for out of disk space
- alert: OutOfDiskSpace
expr: node_filesystem_free_bytes{fstype="ext4"} / node_filesystem_size_bytes{fstype="ext4"} * 100 < 10
for: 5m
labels:
severity: warning
annotations:
summary: "Espace libre de {{ $labels.mountpoint }} sur {{ $labels.instance }} à {{ humanize $value }}%."
# Alert for out of inode space on disk
- alert: OutOfInodes
expr: node_filesystem_files_free{fstype="ext4"} / node_filesystem_files{fstype="ext4"} * 100 < 10
for: 5m
labels:
severity: warning
annotations:
summary: "Presque plus d'inodes disponibles ({{ $value }}% restant) dans {{ $labels.mountpoint }} sur {{ $labels.instance }}."
# Alert for high CPU usage
- alert: CpuUsage
expr: (100 - avg by (instance) (irate(node_cpu_seconds_total{mode="idle"}[5m])) * 100) > 75
for: 10m
labels:
severity: warning
annotations:
summary: "CPU sur {{ $labels.instance }} à {{ humanize $value }}%."
# Check systemd unit (> buster)
- alert: SystemdServiceFailed
expr: node_systemd_unit_state{state="failed"} == 1
for: 10m
labels:
severity: warning
annotations:
summary: "{{ $labels.name }} a échoué sur {{ $labels.instance }}"
# Check load of instance
- alert: LoadUsage
expr: node_load1 > 5
for: 2m
labels:
severity: warning
annotations:
summary: "La charge de {{ $labels.instance }} est à {{ $value }} !"
# Check UPS
- alert: UpsOutputSourceChanged
expr: upsOutputSource != 3
for: 1m
labels:
severity: warning
annotations:
summary: "La source d'alimentation de {{ $labels.instance }} a changé !"
- alert: UpsBatteryStatusWarning
expr: upsBatteryStatus == 3
for: 2m
labels:
severity: warning
annotations:
summary: "L'état de la batterie de {{ $labels.instance }} est faible !"
- alert: UpsBatteryStatusCritical
expr: upsBatteryStatus == 4
for: 10m
labels:
severity: warning
annotations:
summary: "L'état de la batterie de {{ $labels.instance }} est affaibli !"
- alert: UpsHighLoad
expr: upsOutputPercentLoad > 70
for: 5m
labels:
severity: critical
annotations:
summary: "La charge de {{ $labels.instance }} est de {{ $value }}% !"
- alert: UpsWrongInputVoltage
expr: (upsInputVoltage < 210) or (upsInputVoltage > 250)
for: 10m
labels:
severity: warning
annotations:
summary: "La tension d'entrée de {{ $labels.instance }} est de {{ $value }}V."
- alert: UpsWrongOutputVoltage
expr: (upsOutputVoltage < 220) or (upsOutputVoltage > 240)
for: 10m
labels:
severity: warning
annotations:
summary: "La tension de sortie de {{ $labels.instance }} est de {{ $value }}V."
- alert: UpsTimeRemainingWarning
expr: upsEstimatedMinutesRemaining < 8
for: 1m
labels:
severity: warning
annotations:
summary: "L'autonomie restante sur {{ $labels.instance }} est de {{ $value }} min."
- alert: UpsTimeRemainingCritical
expr: upsEstimatedMinutesRemaining < 5
for: 1m
labels:
severity: critical
annotations:
summary: "L'autonomie restante sur {{ $labels.instance }} est de {{ $value }} min."
{% endraw %}

View file

@ -10,19 +10,15 @@
- name: Configure Prometheus
template:
src: prometheus/prometheus.yml.j2
dest: /etc/prometheus/prometheus.yml
mode: 0644
notify: Restart Prometheus
- name: Configure Prometheus alert rules
template:
src: "prometheus/{{ item }}.j2"
src: "{{ item }}.j2"
dest: "/etc/prometheus/{{ item }}"
mode: 0644
notify: Restart Prometheus
owner: prometheus
group: prometheus
mode: u=r,g=r,o=
loop:
- prometheus.yml
- alert.rules.yml
notify: Restart Prometheus
# We don't need to restart Prometheus when updating nodes
- name: Configure Prometheus Federate devices

View file

@ -1,6 +1,10 @@
---
{{ ansible_managed | comment }}
{% macro raw(string) -%}
{{ "{{" }} {{ string }} {{ "}}" }}
{%- endmacro %}
groups:
- name: alert.rules
rules:
@ -11,6 +15,5 @@ groups:
severity: critical
annotations:
summary: >-
Federate : {{ "{{" }} $labels.instance {{ "}}" }} est invisible
depuis plus de 3 minutes !
Invisible depuis plus de 3 minutes
...