UnclePhil / ascaler

ascaler

Current Release: 1.0.0

This project is a complete rewriting of the https://github.com/jcwimer/docker-swarm-autoscaler

Due to the intensive usage of docker service inspect command.

On my cluster(s) the old project was a high cpu user (>15%) and i need to reduce that drastically.

Usage

1. You can deploy prometheus, cadvisor, and ascaler by running docker stack deploy -c swarm-ascaler-stack.yml ascaler from the root of this repo.

• ascaler needs a placement contstraint to deploy to a manager.

2. For services you want to autoscale you will need a deploy label ascaler=true.

deploy:
  labels:
    - "ascaler=true"

This is best paired with resource constraints limits. This is also under the deploy key.

deploy:
  resources:
    reservations:
      cpus: '0.25'
      memory: 512M
    limits:
      cpus: '0.50'

Configuration

Setting	Value	Description
ascaler	true	Required. This enables autoscaling for a service. Anything other than true will not enable it
ascaler.minimum	Integer	Optional. This is the minimum number of replicas wanted for a service. The autoscaler will not downscale below this number
ascaler.maximum	Integer	Optional. This is the maximum number of replicas wanted for a service. The autoscaler will not scale up past this number

Test

You can deploy a test app with the following commands below. Helloworld is initially only 1 replica. The autoscaler will scale to the minimum 3 replicas.

1. docker stack deploy -c swarm-ascaler-stack.yml ascaler
2. docker stack deploy -c helloworld.yml hello