-
Notifications
You must be signed in to change notification settings - Fork 0
Open
Description
description
When using some of the rules, prometheus-server fails to start with the following error message:
Aug 19 00:34:36 server prometheus-server[26365]: level=error ts=2020-08-18T22:34:36.695Z caller=main.go:764 err="error loading config from \"/etc/prometheus/prometheus.yml\": one or more errors occurred while applying the new configuration (--config.file=\"/etc/prometheus/prometheus.yml\")"
Config:
/etc/prometheus/prometheus.yml:
rule_files:
- "rules.d/load-average.yml"
Expected behavior
Prometheus should start without errors.
Proposed fix
Here is a working example of the first rule in the load-average.rules.yml:
groups:
- name: load-average.rules
rules:
- alert: NodeLoadAverage
expr: ((node_load5 / count without (cpu, mode) (node_cpu{mode="system"})) > 1) or (node_load5 / count without (cpu, mode) (node_cpu_seconds_total{mode="system"})) > 1
for: 6h
labels:
severity: warning
priority: P3
paging: false
annotations:
summary: High Load on {{ $labels.instance }} detected
description: "{{ $labels.instance }} Average load is above threshold (current value is: {{ $value }})"
There should be no newlines. I also tried:
foo: |
bla
foo
bar
which also does not work. Only fix I found was to remove the newlines completely.
PS: The disk check for example does work, even with newlines.. I'm not sure what exactly is causing the issue - I just included all your rules at once and then noticed prometheus fails to start.
Metadata
Metadata
Assignees
Labels
No labels