Automating scaling up of nova instances using openstack heat (orchestration) and ceilometer (telementry) services.

1. Utilizing Heat and
Ceilometer to scale up
nova instances
Presented By – Kaustubh S. Dhokte

2. Agenda
 Understanding the problem statement
 Introduction to respective Openstack
Services
 Introduction to HOT
 High level System Functioning
 Low level System Working
 Conclusion

3. Problem Statement:
“Utilizing Heat and Ceilometer to scale
up nova instances”
 Scaling Up
 Horizontally
 Vertically

4. What are these Openstack Services?
Heat – Orchestration Service
Ceilometer – Telementary Service
Nova – Compute Service

5. HOT
 What is it?
 HOT Structure –
 Version
 Parameters
 Resources
 Output
 Intrinsic Functions

7. Intrinsic Functions -
 get_attr
 get_resource
 get_param

8. High Level System Functioning
Ceilometer keeps track of resource
consumption
Ceilometer alarms when CPU Load
increases
Heat hears the alarm
Heat scales up the nova instance
according to the scaling policy

9. Required Resource Types
 OS :: Ceilometer :: Alarm
 OS :: Heat :: ScalingPolicy
 OS :: Heat :: AutoScalingGroup
 OS :: Nova :: Server

10. cpu_alarm_high:
type: OS::Ceilometer::Alarm
properties:
description: scale up if average cpu>80%
for 1 minute
meter_name:cpu_util
statistic:avg
period: 60
threshold: 80
alarm_actions:
- {get_attr:
[nova_instance_scaling_policy, alarm_url]}
comparison_operator:gt

11. nova_instance_scaling_policy:
type: OS::Heat::ScalingPolicy
properties:
adjustment_type:change_in_policy
auto_scaling_group_id:
{get_resource:nova_instance_group}
scaling_adjustment:1
Attibute :- alarm_url(url to handle the alarm)

12. nova_instance_group:
type: OS::Heat::AutoScalingGroup
properties:
min_size : 1
max_size : 3
resource :
type : OS::Nova::Server
properties:
key_name : my_key
image : F18_x64_cfntools
flavour: m1.small

13. Conclusion