Today, we’re excited to announce the general availability of Compute autoscaling V2, now supporting schedule-based scaling. Controlling Compute resource consumption to optimize costs and keeping adequate provisioned capacity for anticipated traffic demand can be challenging. To help address these challenges, autoscaling is introducing new ways to affordably scale predictable load patterns that require scaling on a specific date, time, day, or month.
With your usage and feedback, this release launches significant enhancements that enable the following best-in-class elastic scaling capabilities:
Schedule-based policies: Create up to 50 schedule policies, including one-time scaling, scale out once at a specific date and time, and recurring scaling at regular dates and times.
GPU and DenseIO VM shapes: Expanded autoscaling support for high-performance VM GPU and Dense IO shapes.
Console autoscaling configuration creation wizard: New UI experience that easily and quickly guides you through creating autoscaling configuration and policies.
Support for multiple load balancers: Any autoscaling configuration with an instance pool containing multiple load balancers attached automatically adds backends to each load balancer on a scaling event.
Scale out to 500 instances: Optimized dynamic scale to 500 compute instances per instance pool.
Enable and disable feature: Turn on or off autoscaling configuration or individual scaling policies.
Integration with serverless automation: Cloud native integration to create Event rules on scaling actions or when a pool update is triggered to automatically respond and execute actions, such as Notifications, Functions, or Streaming messages.
With the introduction of these new features, optimizations, and the overhaul of the console UI experience, autoscaling now offers enhanced flexible and agile scaling options. It meets the needs of all workload requirements, provides maximum cost-effectiveness, and offers a fully managed autonomous scaling platform that eliminates the need to monitor capacity.
A new simplified Console user interface has been enhanced with a multi-step walk-through experience. This streamlined interface provides step-by-step guide to quickly configure scaling policies using metrics rules or set a schedule with quick-start cron expressions to easily configure complex scaling schedules.
Scheduled scaling allows you to create a new scheduled-based scaling policy type, where you can set any pre-determined date-time to execute a scaling event. An autoscaling configuration offers the flexibility to set multiple schedule policies with 1-minute granularity between them, each with different target instance pool sizes for scaling out or in. To improve schedule configuration, the integrated scheduler uses a cron scheduler. The cron scheduler enables flexibility to trigger scaling events for recurring predictable traffic loads or one-time-only events when there is known anticipated demand.
The benefits of a schedule-based policy enable the prescale of an instance pool before any spike increase in application load. Typically, this change happens when the traffic demand predictably ramps-up rapidly in a short window of time and there isn’t enough time to launch more Compute instances to handle a surge in new traffic. A metric performance-based policy doesn’t fully meet this requirement, because it scales dynamically and responds to point-in-time changing demand.
Recurring events allow you to schedule scale out capacity horizontally at a specific day/month and time, then scale-in automatically when the extra capacity is no longer needed. The following examples show instances where schedule-based autoscaling is useful:
The web workload every week the traffic pattern increases at 7 a.m. Monday and stays high until 3 p.m.
Dev and test workloads scale-out every weekday in the morning and scale-in in the evening.
One-off events map a known seasonality for a business. The set capacity is prepared in advance ahead of a one-off event, such as the following examples:
The holiday season for retail, such as Cyber Monday
The end of the quarter for financials
The autoscaling configuration details page provides an improved UI to quickly view policies in the same single pane with easier management to add, edit, delete, and the new ability to enable or disable individual policies on demand.
As a leading cloud provider of high-performance Compute instances, we now support autoscaling NVIDIA GPU accelerated VM shapes and Dense IO virtual machine instances in an instance pool. This enterprise-grade elastic scaling matches highly performant application needs.
Autoscaling, coupled with Oracle Cloud Native Services, provides serverless automation support. This event-driven automation allows you to build automation workflows to create serverless Event orchestration from emitted scaling event types. You can set up a serverless rule that triggers when a scaling event happens and configure it to execute actions, such as Notifications, Functions, or Streaming messages. This unified autoscaling and serverless experience provides all the capabilities you need to automate the scaling of your Compute instances and set up post-automation processes for other required infrastructure management.
V2 of Autoscaling brings you the advanced tools to be hands-off and let your Compute environment scale demands be autonomously managed.
Try out this enhanced release of Autoscaling for your workload, if you don’t have an account yet, Oracle Cloud Infrastructure offers US $300 free credit to easily sign up.