openstack/api-site
Code Issues Proposed changes
api-site/firstapp/samples/heat/faafo_autoscaling_workers.yaml
Martin Paulo f799d51ef1 Added an autoscaling heat template 
The template scales the workers both up and down,
dependent on the load across all of the workers. The
orchestration text has been modified to describe
how it all hangs together.

Change-Id: I5959a734ecb21476ab6359cacf49317d370cd0a5
2015-11-25 20:59:52 -06:00

278 lines
8.8 KiB
YAML
Raw Blame History

heat_template_version: 2014-10-16
description: |
A template that starts the faafo application with auto-scaling workers
parameters:
key_name:
type: string
description: Name of an existing keypair to enable SSH access to the instances
default: id_rsa
constraints:
- custom_constraint: nova.keypair
description: Must already exist on your cloud
flavor:
type: string
description: The flavor that the application uses
constraints:
- custom_constraint: nova.flavor
description: Must be a valid flavor provided by your cloud provider.
image_id:
type: string
description: The ID of the image to use to create the instance
constraints:
- custom_constraint: glance.image
description: Must be a valid image on your cloud
period:
type: number
description: The period to use to calculate the ceilometer statistics (in seconds)
default: 60
faafo_source:
type: string
description: The location of the faafo application install script on the Internet
# allows you to clone and play with the faafo code if you like
default: https://git.openstack.org/cgit/openstack/faafo/plain/contrib/install.sh
resources:
api:
type: OS::Neutron::SecurityGroup
properties:
description: "For ssh and http on an api node"
rules: [
{remote_ip_prefix: 0.0.0.0/0,
protocol: tcp,
port_range_min: 22,
port_range_max: 22},
{remote_ip_prefix: 0.0.0.0/0,
protocol: tcp,
port_range_min: 80,
port_range_max: 80},]
worker:
type: OS::Neutron::SecurityGroup
properties:
description: "For ssh on a worker node"
rules: [
{remote_ip_prefix: 0.0.0.0/0,
protocol: tcp,
port_range_min: 22,
port_range_max: 22},]
services:
type: OS::Neutron::SecurityGroup
properties:
description: "For ssh, DB and AMPQ on the services node"
rules: [
{remote_ip_prefix: 0.0.0.0/0,
protocol: tcp,
port_range_min: 80,
port_range_max: 80},
{remote_ip_prefix: 0.0.0.0/0,
protocol: tcp,
port_range_min: 22,
port_range_max: 22},
{remote_ip_prefix: 0.0.0.0/0,
protocol: tcp,
port_range_min: 5672,
port_range_max: 5672,
remote_mode: remote_group_id,
remote_group_id: { get_resource: worker } },
{remote_ip_prefix: 0.0.0.0/0,
protocol: tcp,
port_range_min: 5672,
port_range_max: 5672,
remote_mode: remote_group_id,
remote_group_id: { get_resource: api } },
{remote_ip_prefix: 0.0.0.0/0,
protocol: tcp,
port_range_min: 3306,
port_range_max: 3306,
remote_mode: remote_group_id,
remote_group_id: { get_resource: api } },
]
app_services:
# The database and AMPQ services run on this instance.
type: OS::Nova::Server
properties:
image: { get_param: image_id }
flavor: { get_param: flavor }
key_name: { get_param: key_name }
name: services
security_groups:
- {get_resource: services}
user_data_format: RAW
user_data:
str_replace:
template: |
#!/usr/bin/env bash
curl -L -s faafo_installer | bash -s -- \
-i database -i messaging
wc_notify --data-binary '{"status": "SUCCESS"}'
params:
wc_notify: { get_attr: ['wait_handle', 'curl_cli'] }
faafo_installer: { get_param: faafo_source }
api_instance:
# The web interface runs on this instance
type: OS::Nova::Server
properties:
image: { get_param: image_id }
flavor: { get_param: flavor }
key_name: { get_param: key_name }
name: api
security_groups:
- {get_resource: api}
user_data_format: RAW
user_data:
str_replace:
template: |
#!/usr/bin/env bash
curl -L -s faafo_installer | bash -s -- \
-i faafo -r api -m 'amqp://guest:guest@services_ip:5672/' \
-d 'mysql+pymysql://faafo:password@services_ip:3306/faafo'
wc_notify --data-binary '{"status": "SUCCESS"}'
params:
wc_notify: { get_attr: ['wait_handle', 'curl_cli'] }
services_ip: { get_attr: [app_services, first_address] }
faafo_installer: { get_param: faafo_source }
worker_auto_scaling_group:
# The worker instances are managed by this auto-scaling group
type: OS::Heat::AutoScalingGroup
properties:
resource:
type: OS::Nova::Server
properties:
key_name: { get_param: key_name }
image: { get_param: image_id }
flavor: { get_param: flavor }
# The metadata used for ceilometer monitoring
metadata: {"metering.stack": {get_param: "OS::stack_id"}}
name: worker
security_groups:
- {get_resource: worker}
user_data_format: RAW
user_data:
str_replace:
template: |
#!/usr/bin/env bash
curl -L -s faafo_installer | bash -s -- \
-i faafo -r worker -e 'http://api_1_ip' -m 'amqp://guest:guest@services_ip:5672/'
wc_notify --data-binary '{"status": "SUCCESS"}'
params:
wc_notify: { get_attr: ['wait_handle', 'curl_cli'] }
api_1_ip: { get_attr: [api_instance, first_address] }
services_ip: { get_attr: [app_services, first_address] }
faafo_installer: { get_param: faafo_source }
min_size: 1
desired_capacity: 1
max_size: 3
wait_handle:
type: OS::Heat::WaitConditionHandle
wait_condition:
type: OS::Heat::WaitCondition
depends_on: [ app_services, api_instance, worker_auto_scaling_group ]
properties:
handle: { get_resource: wait_handle }
# All three initial servers clock in when they finish installing their software
count: 3
# 10 minute limit for installation
timeout: 600
scale_up_policy:
type: OS::Heat::ScalingPolicy
properties:
adjustment_type: change_in_capacity
auto_scaling_group_id: {get_resource: worker_auto_scaling_group}
cooldown: { get_param: period }
scaling_adjustment: 1
scale_down_policy:
type: OS::Heat::ScalingPolicy
properties:
adjustment_type: change_in_capacity
auto_scaling_group_id: {get_resource: worker_auto_scaling_group}
cooldown: { get_param: period }
scaling_adjustment: '-1'
cpu_alarm_high:
type: OS::Ceilometer::Alarm
properties:
description: Scale-up if the average CPU > 90% for period seconds
meter_name: cpu_util
statistic: avg
period: { get_param: period }
evaluation_periods: 1
threshold: 90
alarm_actions:
- {get_attr: [scale_up_policy, alarm_url]}
matching_metadata: {'metadata.user_metadata.stack': {get_param: "OS::stack_id"}}
comparison_operator: gt
cpu_alarm_low:
type: OS::Ceilometer::Alarm
properties:
description: Scale-down if the average CPU < 15% for period seconds
meter_name: cpu_util
statistic: avg
period: { get_param: period }
evaluation_periods: 1
threshold: 15
alarm_actions:
- {get_attr: [scale_down_policy, alarm_url]}
matching_metadata: {'metadata.user_metadata.stack': {get_param: "OS::stack_id"}}
comparison_operator: lt
outputs:
api_url:
description: The URL for api server
value:
list_join: ['', ['http://', get_attr: [api_instance, first_address]]]
scale__workers_up_url:
description: >
HTTP POST to this URL webhook to scale up the worker group.
Does not accept request headers or body. Place quotes around the URL.
value: {get_attr: [scale_up_policy, alarm_url]}
scale_workers_down_url:
description: >
HTTP POST to this URL webhook to scale down the worker group.
Does not accept request headers or body. Place quotes around the URL.
value: {get_attr: [scale_down_policy, alarm_url]}
ceilometer_statistics_query:
value:
str_replace:
template: >
ceilometer statistics -m cpu_util -q metadata.user_metadata.stack=stackval -p period -a avg
params:
stackval: { get_param: "OS::stack_id" }
period: { get_param: period }
description: >
This query shows the cpu_util sample statistics of the worker group in this stack.
These statistics trigger the alarms.
ceilometer_sample_query:
value:
str_replace:
template: >
ceilometer sample-list -m cpu_util -q metadata.user_metadata.stack=stackval
params:
stackval: { get_param: "OS::stack_id" }
description: >
This query shows the cpu_util meter samples of the worker group in this stack.
These samples are used to calculate the statistics.
View Git Blame Copy Permalink