CCCHH/ansible-infra
CCCHH/ansible-infra
3
2
Fork 0
Code Issues2 Pull requests Wiki Activity Actions

grafana: differentiate prometheus disk rate alerts by host type

Browse source 
Also only differentiate on the duration not the rate, to not
accidentally exclude slow hard disks.
This commit is contained in:
June 2025-02-06 00:58:55 +01:00
parent 5016407cef
commit a9a5c422bf
Signed by: june
SSH key fingerprint: SHA256:o9EAq4Y9N9K0pBQeBTqhSDrND5E7oB+60ZNx0U1yPe0
1 changed files with 29 additions and 25 deletions

54
resources/chaosknoten/grafana/docker_compose/prometheus_alerts.rules.yaml
View file

@ -114,56 +114,60 @@ groups:
- name: node-exporter-disk
rules:
# Have different disk read and write rate alerts for VMs and physical machines.
- alert: VirtualHostUnusualDiskReadRate
expr: (sum by (instance) (rate(node_disk_read_bytes_total[2m])) / 1024 / 1024 > 50) * on(instance) group_left (nodename) node_uname_info{ype="virtual_machine", nodename=~".+", nodename!="forgejo-actions-runner", nodename!="woodpecker"}
# General high disk read and write rate alerts.
# Excluding: hypervisor hosts, CI hosts
- alert: HostUnusualDiskReadRate
expr: (sum by (instance) (rate(node_disk_read_bytes_total[2m])) / 1024 / 1024 > 50) * on(instance) group_left (nodename) node_uname_info{nodename=~".+", nodename!="forgejo-actions-runner", nodename!="woodpecker", nodename!="chaosknoten"}
for: 5m
labels:
severity: warning
annotations:
summary: Virtual host unusual disk read rate (instance {{ $labels.instance }})
summary: Host unusual disk read rate (instance {{ $labels.instance }})
description: "Disk is probably reading too much data (> 50 MB/s)\n VALUE = {{ $value }}"
- alert: VirtualHostUnusualDiskWriteRate
expr: (sum by (instance) (rate(node_disk_written_bytes_total[2m])) / 1024 / 1024 > 50) * on(instance) group_left (nodename) node_uname_info{type="virtual_machine", nodename=~".+", nodename!="forgejo-actions-runner", nodename!="woodpecker"}
- alert: HostUnusualDiskWriteRate
expr: (sum by (instance) (rate(node_disk_written_bytes_total[2m])) / 1024 / 1024 > 50) * on(instance) group_left (nodename) node_uname_info{nodename=~".+", nodename!="forgejo-actions-runner", nodename!="woodpecker", nodename!="chaosknoten"}
for: 2m
labels:
severity: warning
annotations:
summary: Virtual host unusual disk write rate (instance {{ $labels.instance }})
summary: Host unusual disk write rate (instance {{ $labels.instance }})
description: "Disk is probably writing too much data (> 50 MB/s)\n VALUE = {{ $value }}"
# Some VMs are expected to have high Read / Write rates z.B. CI servers
- alert: VirtualHostUnusualDiskReadRate
expr: (sum by (instance) (rate(node_disk_read_bytes_total[2m])) / 1024 / 1024 > 50) * on(instance) group_left (nodename) node_uname_info{ype="virtual_machine", nodename="forgejo-actions-runner", nodename="woodpecker"}
# CI hosts high disk read and write alerts.
# Longer intervals to account for disk intensive CI tasks.
- alert: CIHostUnusualDiskReadRate
expr: (sum by (instance) (rate(node_disk_read_bytes_total[2m])) / 1024 / 1024 > 50) * on(instance) group_left (nodename) node_uname_info{nodename="forgejo-actions-runner", nodename="woodpecker"}
for: 10m
labels:
severity: warning
annotations:
summary: Virtual host unusual disk read rate for 10 min (instance {{ $labels.instance }})
summary: CI host unusual disk read rate for 10 min (instance {{ $labels.instance }})
description: "Disk is probably reading too much data (> 50 MB/s)\n VALUE = {{ $value }}"
- alert: VirtualHostUnusualDiskWriteRate
expr: (sum by (instance) (rate(node_disk_written_bytes_total[2m])) / 1024 / 1024 > 50) * on(instance) group_left (nodename) node_uname_info{type="virtual_machine", nodename="forgejo-actions-runner", nodename="woodpecker"}
expr: (sum by (instance) (rate(node_disk_written_bytes_total[2m])) / 1024 / 1024 > 50) * on(instance) group_left (nodename) node_uname_info{nodename="forgejo-actions-runner", nodename="woodpecker"}
for: 4m
labels:
severity: warning
annotations:
summary: Virtual host unusual disk write rate for 4 min (instance {{ $labels.instance }})
summary: CI host unusual disk write rate for 4 min (instance {{ $labels.instance }})
description: "Disk is probably writing too much data (> 50 MB/s)\n VALUE = {{ $value }}"
- alert: PhysicalHostUnusualDiskReadRate
expr: (sum by (instance) (rate(node_disk_read_bytes_total[2m])) / 1024 / 1024 > 100) * on(instance) group_left (nodename) node_uname_info{type="physical_machine", nodename=~".+"}
# Hypervisor host high disk read and write alerts.
# Longer intervals to account for disk intensive hypervisor tasks (backups, moving VMs, etc.).
- alert: HypervisorHostUnusualDiskReadRate
expr: (sum by (instance) (rate(node_disk_read_bytes_total[2m])) / 1024 / 1024 > 50) * on(instance) group_left (nodename) node_uname_info{nodename="chaosknoten"}
for: 30m
labels:
severity: warning
annotations:
summary: Hypervisor host unusual disk read rate (instance {{ $labels.instance }})
description: "Disk is probably reading too much data (> 50 MB/s)\n VALUE = {{ $value }}"
- alert: HypervisorHostUnusualDiskWriteRate
expr: (sum by (instance) (rate(node_disk_written_bytes_total[2m])) / 1024 / 1024 > 50) * on(instance) group_left (nodename) node_uname_info{nodename="chaosknoten"}
for: 20m
labels:
severity: warning
annotations:
summary: Physical host unusual disk read rate (instance {{ $labels.instance }})
description: "Disk is probably reading too much data (> 100 MB/s)\n VALUE = {{ $value }}"
- alert: PhysicalHostUnusualDiskWriteRate
expr: (sum by (instance) (rate(node_disk_written_bytes_total[2m])) / 1024 / 1024 > 100) * on(instance) group_left (nodename) node_uname_info{type="physical_machine", nodename=~".+"}
for: 15m
labels:
severity: warning
annotations:
summary: Physical host unusual disk write rate (instance {{ $labels.instance }})
description: "Disk is probably writing too much data (> 100 MB/s)\n VALUE = {{ $value }}"
summary: Hypervisor host unusual disk write rate (instance {{ $labels.instance }})
description: "Disk is probably writing too much data (> 50 MB/s)\n VALUE = {{ $value }}"
# Please add ignored mountpoints in node_exporter parameters like
# "--collector.filesystem.ignored-mount-points=^/(sys|proc|dev|run)($|/)".
# Same rule using "node_filesystem_free_bytes" will fire when disk fills for non-root users.