AutoScaling API

Introduction

Auoscaling allows you to dynamically adjust the number of Instances within a given Instance group based on defined scaling policies. Scaling actions (scale up or down) are triggered when the monitored metric exceeds the configured thresholds from your policies. Autoscaling relies on the Cockpit product to gather the Instance metrics, such as RAM or bandwidth usage. The Load Balancer of your choice, which handles traffic for the given Instance group, is automatically configured according to the scaling action so that connections are properly distributed among the set of Instances.

The purpose of this product is to maintain optimal application performance and cost efficiency by scaling your Instance Group up during peak traffic and scaling it down when demand decreases. This ensures that applications have the necessary resources to handle varying loads without manual intervention, providing high availability and fault tolerance. Autoscaling is particularly useful for dynamic workloads, enabling businesses to optimize resource usage and reduce operational overhead.

Concepts

Refer to our dedicated concepts page to find definitions of all autoscaling-related terminology.

Quickstart

1. Configure your environment variables.

   Note

   This is an optional step that seeks to simplify your usage of the Autoscaling API.

   export SCW_SECRET_KEY="<API secret key>"
   export SCW_DEFAULT_ZONE="<Scaleway default Availability Zone>"
   export SCW_PROJECT_ID="<Scaleway Project ID>"
   

2. Create an Instance Template: run the following command to create an Instance template. The created template will be used during scaling to start new Instances, according to the given settings.

   curl -X POST \
     -H "X-Auth-Token: $SCW_SECRET_KEY" \
     -H "Content-Type: application/json" \
     "https://api.scaleway.com/autoscaling/v1alpha1/zones/$SCW_DEFAULT_ZONE/instance-templates" \
     -d '{
       "name": "my_template",
       "commercial_type": "PLAY2-NANO",
       "volumes": {
           "0": {
               "name": "boot volume",
               "from_snapshot": {
                   "snapshot_id": "<SNAPSHOT_ID>"
               },
               "volume_type": "sbs",
               "boot": true
           }
       },
       "private_network_ids": [
           "<PRIVATE_NETWORK_ID>"
       ],
       "project_id": "'"$SCW_PROJECT_ID"'"
     }'
   

3. Create an Instance Group: run the following command to create an Instance group. You must have already configured a Load Balancer, so that the Instance group can update its backend servers list during scaling actions. You can specify the Private Network to be used for traffic between your Load Balancer and the Instances. If specified, the given Private Network must be already attached to the Load Balancer.

   curl -X POST \
     -H "X-Auth-Token: $SCW_SECRET_KEY" \
     -H "Content-Type: application/json" \
     "https://api.scaleway.com/autoscaling/v1alpha1/zones/$SCW_DEFAULT_ZONE/instance-groups" \
     -d '{
       "project_id": "'"$SCW_PROJECT_ID"'",
       "name": "my_instance_group",
       "template_id": "<AUTOSCALING_TEMPLATE_ID>",
       "capacity": {
         "max_replicas": 5,
         "min_replicas": 1,
         "cooldown_delay": "300s"
       },
       "loadbalancer":{
           "id": "<LOAD_BALANCER_ID>",
           "backend_ids": ["<BACKEND_ID>"],
           "private_network_id": "<PRIVATE_NETWORK_ID>"
       }
     }'
   

4. Create Instance Policies: run the following command to create Instance policies. Instance policies are a set of rules which define criteria for scaling actions. Create at least 2 policies for both scaling up and down actions.

   curl -X POST \
     -H "X-Auth-Token: $SCW_SECRET_KEY" \
     -H "Content-Type: application/json" \
     "https://api.scaleway.com/autoscaling/v1alpha1/zones/$SCW_DEFAULT_ZONE/instance-policies" \
     -d '{
       "name": "my-policy-scale-up",
       "metric": {
         "name": "cpu scale up",
         "managed_metric": "managed_metric_instance_cpu",
         "operator": "operator_greater_than",
         "aggregate": "aggregate_average",
         "sampling_range_min": 5,
         "threshold": 70
       },
       "action": "scale_up",
       "type": "flat_count",
       "value": 1,
       "priority": 1,
       "instance_group_id": "<AUTOSCALING_INSTANCE_GROUP_ID>"
     }'
   curl -X POST \
     -H "X-Auth-Token: $SCW_SECRET_KEY" \
     -H "Content-Type: application/json" \
     "https://api.scaleway.com/autoscaling/v1alpha1/zones/$SCW_DEFAULT_ZONE/instance-policies" \
     -d '{
       "name": "my-policy-scale-down",
       "metric": {
         "name": "cpu scale down",
         "managed_metric": "managed_metric_instance_cpu",
         "operator": "operator_less_than",
         "aggregate": "aggregate_average",
         "sampling_range_min": 5,
         "threshold": 40
       },
       "action": "scale_down",
       "type": "flat_count",
       "value": 1,
       "priority": 2,
       "instance_group_id": "<AUTOSCALING_INSTANCE_GROUP_ID>"
     }'
   

5. Get a list of the scaling events: run the following command to get a list of all the scaling events which occurred for the given Instance group:

   curl -X GET \
     -H "Content-Type: application/json" \
     -H "X-Auth-Token: $SCW_SECRET_KEY" \
     "https://api.scaleway.com/autoscaling/v1alpha1/zones/$SCW_DEFAULT_ZONE/instance-groups/<INSTANCE_GROUP_ID>/events"
   

6. Delete your Instance Group: run the following command to delete an Instance group. Ensure that you replace <INSTANCE_GROUP_ID> in the URL with the ID of the Instance group you want to delete.

   curl -X DELETE \
     -H "Content-Type: application/json" \
     -H "X-Auth-Token: $SCW_SECRET_KEY" \
     "https://api.scaleway.com/autoscaling/v1alpha1/zones/$SCW_DEFAULT_ZONE/instance-groups/<INSTANCE_GROUP_ID>"
   

Technical information

Autoscaling relies on the Cockpit product to gather metrics and take scaling actions based on defined policies. Scaling can thus occur every 5 minutes, which is the time to re-fetch metrics from Cockpit after a previous iteration.

Availability Zones

The Scaleway Autoscaling API is a zoned API, meaning that each call must specify in its path parameters the Availability Zone for the resources concerned by the call.

The following Availability Zones are available for Autoscaling:

Technical limitations

• Autoscaling does not replicate changes to existing Instances after an update of an Instance template update. If you need to refresh all existing Instances so that they comply with the new template, you must terminate them so that Autoscaling can create new Instances based on the updated template.
• Custom metrics are not supported yet. You can only set scaling policies according to pre-defined "managed metrics".
• Autoscaling does not handle Load Balancer and its associated backend creation. It only manages backend server updates for the given existing backend.
• Autoscaling relies on the Instance status to determine whether it is healthy or not, not on the Load Balancer health check mechanism.

Going further

For more help using Scaleway Autoscaling, check out the following resources:

• The #autoscaling channel on our Slack Community
• Our support ticketing system.

Instance Groups

Instance Templates

Instance Policies