k-space
/
kube
10
1
Fork 1
Code Issues 17 Pull Requests Projects Activity

prometheus-operator: Fix node exporter formatting

This commit is contained in:
Lauri Võsandi 2022-09-12 07:44:06 +03:00
parent b5b5cd038e
commit 7ae75f2f1c
1 changed files with 342 additions and 341 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

11
prometheus-operator/node-exporter.yml
View File

@ -1,3 +1,4 @@
---
apiVersion: monitoring.coreos.com/v1
kind: Probe
metadata:
@ -243,7 +244,7 @@ spec:
severity: info
annotations:
summary: Host EDAC Correctable Errors detected (instance {{ $labels.instance }})
description: {{ $labels.instance }} has had {{ printf "%.0f" $value }} correctable memory errors reported by EDAC in the last 5 minutes.
description: "Host {{ $labels.instance }} has had {{ printf \"%.0f\" $value }} correctable memory errors reported by EDAC in the last 5 minutes.\n VALUE = {{ $value }}\n LABELS = {{ $labels }}"
- alert: HostEdacUncorrectableErrorsDetected
expr: node_edac_uncorrectable_errors_total > 0
for: 0m
@ -251,7 +252,7 @@ spec:
severity: warning
annotations:
summary: Host EDAC Uncorrectable Errors detected (instance {{ $labels.instance }})
description: {{ $labels.instance }} has had {{ printf "%.0f" $value }} uncorrectable memory errors reported by EDAC in the last 5 minutes.
description: "Host {{ $labels.instance }} has had {{ printf \"%.0f\" $value }} uncorrectable memory errors reported by EDAC in the last 5 minutes.\n VALUE = {{ $value }}\n LABELS = {{ $labels }}"
- alert: HostNetworkReceiveErrors
expr: rate(node_network_receive_errs_total[2m]) / rate(node_network_receive_packets_total[2m]) > 0.01
for: 2m
@ -259,7 +260,7 @@ spec:
severity: warning
annotations:
summary: Host Network Receive Errors (instance {{ $labels.instance }})
description: {{ $labels.instance }} interface {{ $labels.device }} has encountered {{ printf "%.0f" $value }} receive errors in the last five minutes.
description: "Host {{ $labels.instance }} interface {{ $labels.device }} has encountered {{ printf \"%.0f\" $value }} receive errors in the last two minutes.\n VALUE = {{ $value }}\n LABELS = {{ $labels }}"
- alert: HostNetworkTransmitErrors
expr: rate(node_network_transmit_errs_total[2m]) / rate(node_network_transmit_packets_total[2m]) > 0.01
for: 2m
@ -267,7 +268,7 @@ spec:
severity: warning
annotations:
summary: Host Network Transmit Errors (instance {{ $labels.instance }})
description: {{ $labels.instance }} interface {{ $labels.device }} has encountered {{ printf "%.0f" $value }} transmit errors in the last five minutes.
description: "Host {{ $labels.instance }} interface {{ $labels.device }} has encountered {{ printf \"%.0f\" $value }} transmit errors in the last two minutes.\n VALUE = {{ $value }}\n LABELS = {{ $labels }}"
- alert: HostNetworkInterfaceSaturated
expr: (rate(node_network_receive_bytes_total{device!~"^tap.*"}[1m]) + rate(node_network_transmit_bytes_total{device!~"^tap.*"}[1m])) / node_network_speed_bytes{device!~"^tap.*"} > 0.8
for: 1m
@ -275,7 +276,7 @@ spec:
severity: warning
annotations:
summary: Host Network Interface Saturated (instance {{ $labels.instance }})
description: The network interface "{{ $labels.interface }}" on "{{ $labels.instance }}" is getting overloaded.
description: "The network interface {{ $labels.interface }} on {{ $labels.instance }} is getting overloaded."
- alert: HostNetworkBondDegraded
expr: node_bonding_active != node_bonding_slaves {master=~"bond.*"}
for: 2m