Get Started with Event-based Scaling

This topic describes how to quickly get started with event-based scaling for one-time event.

Overview

Installing the Smart Scaler Agent installs and configures the Inference Agent and the Event Autoscaler on your cluster.

The Event Autoscaler proactively scales pods ahead of scheduled events for seamless operations.

Prerequisites

• As described in Sign Up and Quick Start, you must have signed up and installed the Smart Scaler Agent.
• Installing the Smart Scaler Agent configures the Event AutoScaler through the ss-appconfig-values.yaml file on your cluster.
• By default, the in-cluster scaling approach for event infrastructure scaling is enabled. We also support GitOps scaling approach, which requires additional configuration.

Create an Event for Scaling on the Management Console

warning

For event scaling with Smart Scaler agent versions 2.9.28 or earlier, each application deployment in a namespace must have its own configured HPA. Without an individual HPA, event scaling fails.

However, starting with Smart Scaler agent versions 2.9.29 and later, event scaling no longer requires an individual HPA for each application deployment.

1. On the SaaS management console, go to Events on the left sidebar.

2. On the Events page, click the Add Event button towards the top-right of the page.

3. On the Add Event pane, enter a name for the scheduled event you want to scale in the Event Name text box. For example, Acme Fitness Store Sale can be an event that you want to scale.

   .

4. (Optional) Add tags to identify this event later in the Tags text box.

5. Under Event Details:

   1. Enter the scheduled event's start date, start time, end date, and end time in the corresponding text boxes.

      info

      The start date should be the current/future date and the start time should at least 15 minutes in the future.

   2. Select Does Not Repeat (default) for one-time occurring event. The other option is Recurring.

   3. In the Cooldown Period text box, by default 1 hour is selected. You can set the value in hours and in minutes.

      info

      This is a time period during which the auto-scaling system pauses further scaling actions after a scaling event has occurred. It helps prevent rapid and frequent scaling adjustments in response to short-term fluctuations or noise in the traffic.

   4. A green color is by set in the Event Color drop-down list. Change the color if you want to.

   5. In the Event Description text box, describe the event about its purpose or significance.

6. Under Application Details:

   1. Select your application from the drop-down list.

   2. Choose the application location that populates below Choose Locations. You can choose multiple locations if the application is in multiple clusters.

   3. Select the scaling factor which is the projected increase in users. It must be greater than 0 for the location you have chosen.

      You can select any one of the provided scaling factors or add a custom scaling factor that is greater than 0.

7. Click Add Event. The event you just created appears on the Events page.

   .

This creates your first event for the Event Autoscaler to scale its pods for seamless operations proactively!

Scale Planned Events with GitOps Approach

By default, the In-Cluster scaling approach is enabled on the Event Smart Scaler.

To configure scaling with GitOps approach on your cluster, see Event Scaling Using GitOps Approach.

To use the GitOps approach, see Scaling of Planned Events.

Learn More

For more event operations and create a recurring event, see Event Operations.

To gain more flexibility over your application, see Define the Application using Helm.