linux-formula/linux/meta/heka.yml

{%- from "linux/map.jinja" import system with context %}
log_collector:
  decoder:
    system:
      engine: sandbox
      module_file: /usr/share/lma_collector/decoders/generic_syslog.lua
      module_dir: /usr/share/lma_collector/common;/usr/share/heka/lua_modules
      config:
        syslog_pattern: '<%PRI%>%TIMESTAMP% %HOSTNAME% %syslogtag%%msg:::sp-if-no-1st-sp%%msg%\n'
        fallback_syslog_pattern: '%TIMESTAMP% %HOSTNAME% %syslogtag%%msg:::sp-if-no-1st-sp%%msg%\n'
        {%- if system.timezone is defined %}
        tz: "{{ system.timezone }}"
        {%- endif %}
  input:
    linux_log_stream:
      engine: logstreamer
      log_directory: "/var/log"
      file_match: '(?P<Service>daemon\.log|cron\.log|haproxy\.log|kern\.log|auth\.log|syslog|messages|debug)'
      differentiator: [ 'system.', 'Service' ]
      decoder: "system_decoder"
      splitter: "TokenSplitter"
  filter:
    linux_hdd_errors:
      engine: sandbox
      module_file: /usr/share/lma_collector/filters/hdd_errors_counter.lua
      module_dir: /usr/share/lma_collector/common;/usr/share/heka/lua_modules
      preserve_data: false
      message_matcher: "Type == 'log' && Logger == 'system.kern'"
      ticker_interval: 10
      config:
        grace_interval: 10
        patterns: "/error%s.+([sv]d[a-z][a-z]?)%d?/ /([sv]d[a-z][a-z]?)%d?.+%serror/"
        hostname: '{{ grains.host }}'
    linux_logs_counter:
      engine: sandbox
      module_file: /usr/share/lma_collector/filters/logs_counter.lua
      module_dir: /usr/share/lma_collector/common;/usr/share/heka/lua_modules
      preserve_data: true
      message_matcher: "Type == 'log' && Logger =~ /^openstack\\\\./"
      ticker_interval: 1
      config:
        interval: 60
        grace_interval: 30
        hostname: '{{ grains.host }}'
metric_collector:
  trigger:
    linux_system_cpu_critical:
      description: 'The CPU usage is too high.'
      severity: critical
      rules:
      - metric: cpu_wait
        relational_operator: '>='
        threshold: 35
        window: 120
        periods: 0
        function: avg
      - metric: cpu_idle
        relational_operator: <=
        threshold: 5
        window: 120
        function: avg
    linux_system_cpu_warning:
      description: 'The CPU wait times are high.'
      severity: critical
      rules:
      - metric: cpu_wait
        relational_operator: '>='
        threshold: 15
        window: 120
        periods: 0
        function: avg
  alarm:
    linux_system_cpu:
      notifications: False
      alerting: True
      triggers:
      - linux_system_cpu_warning
      - linux_system_cpu_critical