heat/meta/prometheus.yml

{%- if pillar.heat.server is defined %}

{%- from "heat/map.jinja" import server with context %}
{%- if server.get('enabled', False) %}
{% raw %}
server:
  alert:
    HeatAPIDown:
      if: >-
        max(openstack_api_check_status{service=~"heat.+"}) by (service) == 0
      for: 2m
      labels:
        severity: down
        service: "{{ $labels.service }}"
      annotations:
        summary: "Endpoint check for '{{ $labels.service }}' is down"
        description: >-
            Endpoint check for '{{ $labels.service }}' is down for 2 minutes
    HeatErrorLogsTooHigh:
{%- endraw %}
      {%- set log_threshold = prometheus_server.get('alert', {}).get('HeatErrorLogsTooHigh', {}).get('var', {}).get('threshold', 0.2 ) %}
      if: >-
        sum(rate(log_messages{service="heat",level=~"error|emergency|fatal"}[5m])) without (level) > {{ log_threshold }}
{%- raw %}
      labels:
        severity: warning
        service: "{{ $labels.service }}"
      annotations:
        summary: 'Too many errors in {{ $labels.service }} logs'
        description: 'The rate of errors in {{ $labels.service }} logs over the last 5 minutes is too high on node {{ $labels.host }} (current value={{ $value }}, threshold={%- endraw %}{{ log_threshold }}).'
{%- endif %}
{%- endif %}