Ask Your Question
0

Adding meter for 'memory.usage'

asked 2017-09-18 04:01:47 -0500

CDOT gravatar image

updated 2017-09-18 04:02:41 -0500

I followed this link and launched auto-scaling instances that scale based on average CPU utilization. From here, I gather that memory.usage meter is similar to cpu_util meter; the difference being that the former's unit is in MB whereas the latter's unit is %-age.

I would like to add a new alarm that triggers when memory.usage > threshold_mb. The relevant portion of my stack template is :

ram_alarm_high:
    type: OS::Ceilometer::Alarm
    properties:
      meter_name: memory.usage
      statistic: avg
      period: 60
      evaluation_periods: 1
      threshold: { get_param: threshold_mb }
      alarm_actions: [{ get_param: action_url }]
      ok_actions: [{ get_param: ok_url }]
      comparison_operator: gt

Now, I would like to make necessary changes to my /etc/ceilometer/pipeline.yaml so that Memory usage data is collected once every 60 seconds (as specified in the template yaml). How do I proceed ?

edit retag flag offensive close merge delete

1 answer

Sort by ยป oldest newest most voted
1

answered 2017-09-18 05:40:05 -0500

Shaoman Zhang gravatar image

updated 2017-09-18 05:41:46 -0500

Modify the meter_source interval in compute node, like this:

sources:
    - name: meter_source
      interval: 60
      meters:
          - "*"
      sinks:
          - meter_sink
edit flag offensive delete link more

Comments

Many thanks.

CDOT gravatar imageCDOT ( 2017-09-19 00:55:32 -0500 )edit

Your Answer

Please start posting anonymously - your entry will be published after you log in or create a new account.

Add Answer

Get to know Ask OpenStack

Resources for moderators

Question Tools

1 follower

Stats

Asked: 2017-09-18 04:01:47 -0500

Seen: 199 times

Last updated: Sep 18 '17