46 lines
1.4 KiB
YAML
46 lines
1.4 KiB
YAML
|
# Copyright (c) 2018-2020, diyIT.org
|
||
|
|
# All rights reserved.
|
||
|
|
#
|
||
|
|
# BSD 2-Clause License ("Simplified BSD License" or "FreeBSD License")
|
||
|
|
# https://diyit.org/license/
|
||
|
|
#
|
||
|
|
#
|
||
|
|
|
||
|
|
groups:
|
||
|
|
- name: Hardware alerts
|
||
|
|
rules:
|
||
|
|
- alert: Node_down
|
||
|
|
expr: up{job="node_exporter"} == 0
|
||
|
|
for: 1m
|
||
|
|
labels:
|
||
|
|
severity: warning
|
||
|
|
annotations:
|
||
|
|
summary: "Node {{ $labels.instance }} is down"
|
||
|
|
description: "Failed to scrape {{ $labels.job }} on {{ $labels.instance }} for more than 1 minute. Node seems down."
|
||
|
|
- alert: High_cpu_util
|
||
|
|
expr: node_load5{job="node_exporter"} > 6
|
||
|
|
for: 2m
|
||
|
|
labels:
|
||
|
|
severity: warning
|
||
|
|
annotations:
|
||
|
|
summary: "CPU {{ $labels.instance }} is high"
|
||
|
|
description: "{{ $labels.job }} on {{ $labels.instance }} loaded more than 6 for more than 2 minute."
|
||
|
|
- alert: High_disk_util
|
||
|
|
expr: gstat_percent_busy{job="gstat"} > 90
|
||
|
|
for: 3m
|
||
|
|
labels:
|
||
|
|
severity: warning
|
||
|
|
annotations:
|
||
|
|
summary: "Disk {{ $labels.instance }} is loaded"
|
||
|
|
description: "{{ $labels.job }} on {{ $labels.instance }} loaded more than 90% for more than 3 minute."
|
||
|
|
- name: Probe alerts
|
||
|
|
rules:
|
||
|
|
- alert: Site_down
|
||
|
|
expr: probe_success{job="web-diyit"} == 0
|
||
|
|
for: 1m
|
||
|
|
labels:
|
||
|
|
severity: warning
|
||
|
|
annotations:
|
||
|
|
summary: "Site {{ $labels.instance }} is down"
|
||
|
|
description: "Failed to scrape {{ $labels.job }} on {{ $labels.instance }} for more than 1 minute. Site seems down."
|