Containers autoscaling reference

Benefits of autoscalingLink to this anchor

Autoscaling offers several benefits, including improved responsiveness and cost efficiency. By automatically adjusting the number of instances of your resource based on current demand, you ensure that your applications can handle varying loads without manual intervention. This not only enhances user experience by maintaining performance levels but also helps in reducing costs by only using resources when necessary. Additionally, autoscaling helps in optimal resource utilization, minimizing the risk of performance degradation during peak times.

Minimum and maximum scalesLink to this anchor

Minimum scale (min-scale)Link to this anchor

This parameter sets the lowest value your resource is allowed to scale down to:

• If you set a value of 0, all instances of your resource will be terminated after 15 minutes of inactivity.

• If you set a value of 1 or more, the corresponding number of instances of your resource will remain available at all time.

Customizing the minimum scale for Serverless can help ensure that an instance remains pre-allocated and ready to handle requests, reducing delays associated with cold starts. However, this setting also impacts the costs of your Serverless Container.

Maximum scale (max-scale)Link to this anchor

This parameter sets the maximum number of instances of your resource. You should adjust it based on your resource’s traffic spikes, keeping in mind that you may wish to limit the max scale to manage costs effectively.

When the maximum scale is reached, new requests are queued for processing. When the queue is full, the resource returns 503 errors for requests received beyond this point.

Autoscaler behaviorLink to this anchor

The autoscaler decides to start new instances when:

• the existing instances are no longer able to handle the load because they are busy responding to other ongoing requests. By default, this happens if an instance is already processing 80 requests (max_concurrency = 80).

• our system detects an unusual number of requests. In this case, some instances may be started in anticipation to avoid a potential cold start.

The same autoscaler decides to remove instances when:

• no more requests are being processed. If even a single request is being processed (or detected as being processed), then the autoscaler will not be able to remove this instance. The system also prioritizes instances with the fewest ongoing requests, or if very few requests are being sent, it tries to select a particular instance to shut down the others, and therefore scale down.
• an instance has not responded to a request for more than 15 minutes of inactivity. The instance is only shut down after this interval, once again to absorb any potential new peaks and thus avoid the cold start. These 15 minutes of inactivity are not configurable.