Scaling

LayerOps provides flexible autoscaling options that allow you to fine-tune your infrastructure's resource allocation to meet your specific requirements. The autoscaling process can be configured to operate within a range of 0 to N instances, providing tailored control over resource management.

  • Minimum and maximum configuration: With LayerOps, you set a range for the number of instances in your instance pool, defining a minimum and maximum limit. For instance, you can specify that your instance pool should have a minimum of 2 instances and a maximum of 10 instances.
  • Dynamic scaling: Autoscaling at the instance level is responsive to service demand. If there is no service to host, LayerOps can be configured to scale down to zero instances, saving resources and costs. As services require deployment and available instances reach their maximum capacity, LayerOps triggers a scale out operation to create additional instances for hosting new services.
  • Idle instance: LayerOps efficiently manages instances. When an instance no longer hosts any deployed services, it will initiate a scale-down action, reclaiming resources to prevent unnecessary costs.
  • Cooldown: Represents the time interval that LayerOps will wait after a scaling operation, such as scale out or scale down, before initiating another scaling operation. It ensures that scaling adjustments are made strategically, preventing excessive or rapid changes to your infrastructure.

LayerOps also offers autoscaling at the service level

Auto replacement

Auto Replace is a feature of LayerOps that works seamlessly with Auto Scaling. It plays a crucial role in maintaining high availability by automatically replacing instances that have encountered issues, such as network problems or provider-related disruptions.

When Auto Scaling is enabled for your instance pool, Auto Replace is also automatically activated.

If an instance is found to be unresponsive or experiencing issues, LayerOps detects this situation. It identifies instances that are no longer functioning correctly and initiates the process of both terminating the faulty instance and creating a new one in parallel.