FUSAKLA

$ promruval validate --config-file validation.yaml rules.yaml
Validation rules used:
  
  check-severity-label:
    - has labels: `severity`
    - label `severity` has one of the allowed values: `info`,`warning`,`critical`
  
  check-playbook-annotation:
    - has all of these annotations: `playbook`,`title`
    - Annotation `playbook` is a valid URL and does not return HTTP status 404

FUSAKLA

groups:
   - name: test-group
     rules:
       - alert: TrafficDifference1w
         expr: |
          abs(
            rate(http_requests_total{dc="us-west"}[5m])
              - 
            rate(http_requests_total{dc="us-west"}[5m] offset 7d)
          ) > 1000

FUSAKLA

 validationRules:
   - name: check-severity-label
     scope: Alert # Defines which rule types should be validated by this rule.
     validations:
       - type: hasLabels # Name of the validation check.
         params: # Each validation check has its own params depending on the type.
           labels: [ "severity" ]
       - type: labelHasAllowedValue
         params:
           label: "severity"

FUSAKLA

#!/usr/bin/python

import logging
import os
from datetime import datetime, timedelta

import django

django.setup()

FUSAKLA

Intro

So your dashboard is slow? We are doing our best but maybe it's time to think if the issue isn't on your side.

Some of our dashboards looks like https://www.theworldsworstwebsiteever.com/. Are you sure you really need to see all the time all the data? Or do you really need to see data for last week or month all the time? How often do you use filtering by some of those variables in your dashboard?

Let's not end up as your mother saying:

But.. but.. I might need this in future so let's leave this icon on the Desktop.

FUSAKLA

apiVersion: v1
kind: ConfigMap
metadata:
name: alertmanager-conf
labels:
    name: alertmanager-conf
namespace: monitoring
data:
  alertmanager.yml: |-
      global:

FUSAKLA

apiVersion: v1
data:
  prometheus.rules: |
    # TODO add rules
  prometheus.alerts: |
    # host rules
    ALERT high_node_load
      IF node_load1 > 20
      FOR 10s
      LABELS { severity = "critical" }