Article written by

Mehdi Abaakouk (sileht) is a senior python developer, he mainly works on Ceilometer as Core Developer. When he doesn't work on Openstack, he does DevOps stuffs for the non-profit ISP or dance on crazy swing rhythm.

11 Responses

  1. Autoscaling with Heat and Ceilometer | OpenStackうぉっち

    […] 投稿日: 2013/08/29 投稿者: ntamaoki Like AWS CloudFormation, Heat allows to create auto scaling stacks. In order to do this, some metrics need to be retrieved from your VM and some actions need to be triggered when a specified event occurs on these metrics. These actions are usually upscaling (create some new VMs) or downscaling (destroy some VMs). In OpenStack Grizzly, a simplistic system was… Read more ⇒ […]

  2. DeeJay1
    DeeJay1 at |

    Looks nice! Any hints on how to integrate external monitoring tools with it? Eg pushing alarms into ceilometers DB from outside.

  3. OpenStack Community Weekly Newsletter (Aug 23-30) » The OpenStack Blog

    […] Autoscaling with Heat and Ceilometer […]

  4. History, objectives and limits of the Ceilometer project

    […] more details on this, I recommend you to read the blog post by my colleague Mehdi Abaakouk “Autoscaling with Heat and Ceilometer”. A particular attention should be given to the section “Some notes about deploying […]

  5. OpenStack社区周报(8.26 – 9.3) | UnitedStack Inc.

    […] Heat与Ceilometer实现自动缩放的云 […]

  6. OpenStack Community Weekly Newsletter (Aug 23-30) | Keep America At Work

    […] Autoscaling with Heat and Ceilometer […]

  7. Greg Waines
    Greg Waines at |

    Actually I have a general question about alarming in OpenStack.

    Service ‘ceilometer-alarm-singleton’ generates ‘Threshold’ Alarms based on threshold crossings of the meters it is monitoring.

    And service ‘ceilometer-alarm-notifier’ listens to the RPC bus for Alarm Notification events and executes the defined triggered action(s) for any alarms.

    However, a System has more than just meter-based threshold alarms. E.g. a Compute Node Failure, a physical interface failure, etc. .
    Is there a “GENERAL” RPC Bus for Alarm Notification Events in OpenStack ?
    Is it the one used by Ceilometer ?
    Is the intent that Ceilometer is also collecting “ALL” Alarms, storing them in DB, and supporting ‘publishers’ for Alarms (e.g. an SNMP Trap publisher )

    Comments ?

  8. OpenStack社区周报(8.26 – 9.3) | UnitedStack

    […] Heat与Ceilometer实现自动缩放的云 […]

  9. Skylab
    Skylab at |

    Actually I am planning start implement the metering for autoscaling .

    I’ve created stack with lb_members also. Whenever I stress the CPU utilization more than 90%, But still VM scaling is not happening .

    Moreover I see ceilometer alarm-list insufficient data.

    [root@icmcs heat-templates]# ceilometer alarm-list
    | Alarm ID | Name | State | Severity | Enabled | Continuous | Alarm condition | Time constraints |
    | 577063fd-8ae3-4183-95b8-51c94d8e0c7d | demo-autoscaling-cpu_alarm_high-jsjsmtwv2yei | insufficient data | low | True | True | cpu_util > 50.0 during 1 x 60s | None |
    | 75550108-2062-4ca6-afca-bda0886bd524 | lb-members-lb_connection_rate_lo-2jxs347np3ad | insufficient data | low | True | True | 2.0 during 3 x 60s | None |
    | f7d1d20c-9849-42be-9620-8bbcbaba2488 | demo-autoscaling-cpu_alarm_low-yqxw7o5sonbg | insufficient data | low | True | True | cpu_util < 15.0 during 1 x 600s | None |

    [root@icmcs heat-templates]# ceilometer sample-list –meter cpu_util
    | Resource ID | Name | Type | Volume | Unit | Timestamp |

    Can you have chance help on this?. Why I am facing thi issue?. If I need do any modification on .conf please let me know. thanks.

Comments are closed.