People often make a lot of purchases during holiday season, intending to buy gifts for their loved ones. This means more people are using online websites at this time more than at any other point in a given year. Often, this leads to websites crashing because they are not set up to handle this kind of load, as this spike in visitor traffic is an anomaly.
That’s where IBM Auto Scaling comes into play. Businesses can use the IBM cloud to scale with
Auto Scaling, and test loads using the Load Impact service.
What is IBM Auto Scaling?
The Auto Scaling for IBM Cloud service allows an automatic increase or decrease the computing capacity of the application. You can define an Auto Scaling policy, which helps your web application to adjust dynamically by automatically adding or removing resources basis the current workload requirement. You can also view your scaling history and metric statistics, so you can keep track of your dynamic scaling, and make modifications to the scaling policy and metrics as and when necessary.
What is Load Impact Service?
Load Impact is a SaaS performance testing solution. Overall, it helps to reduce development costs and increase revenue. Load Impact also provides an automated results analysis which simplifies interpretation of test results, which helps saves time and allows developers to work on solving issues swiftly. Load Impact can also be used to run large load tests in the cloud for the later developmental stages. Thus, developers are able to fix performance issues, if any, before launch.
This has wide ranging applications in e-commerce and online retail websites. It is perfect for those who are likely to experience fluctuating traffic and need to be able to keep up with an unexpected
influx of website visitors. Businesses may lose out on a lot of visitors if their backend structure cannot accommodate the increased load.
This is likely to lead to loss in revenue as well as customer satisfaction, as visitors may not wait for the website to get back on track and may even go to a competitor’s website. It is important to avoid this, because along with current revenue loss, it may even lead to a long term loss of an otherwise returning customer.
Spikes and unexpected traffic can lead to increased sales and IBM Auto Scaling will ensure that businesses do not lose out of revenue by handling dynamic adjustments in the background. This ensures that your web application will be able to deal with a fluctuating influx of visitors from the backend itself according to your previously specified scaling policy.
Development and Process Description
This requires you to be signed up for an IBM Cloud Account, and your web application needs to be hosted on IBM Cloud as well. After this, all you need to have is a provisional Auto Scaling service on IBM Cloud, and a provisional Load Impact service (at additional cost).
Once you have all that in place, let’s go on to see how you can use Auto Scaling and Load Impact services in order to scale and test your web application.
Attach the IBM Auto Scaling Service to your application
Navigate to your web app and click the Overview tab
Click Create connection and connect the Load Impact service that you provisioned.
As soon as you connect, you see an option to restage your application. Click Restage so that the service becomes available for your web application.
Set up the IBM Auto Scaling Service according to your requirements
Navigate to the Auto Scaling service you recently attached to your web app. Select the option to
view the policy configuration, and create an auto scaling policy.
In this configuration, you can scale your app using the following four metrics:
4. Average time
Different auto scaling policy configurations are available based on the metric selected. This example includes a policy configured for the memory to scale out when the upper threshold goes above 30 percent of the memory and to scale in when the lower threshold goes below 10 percent of the memory. You can add multiple rules, but make sure you do not add conflicting rules which might cause an issue.
If you use advanced policies, you can set application characteristics for the following four metrics:
1. Statistic window
2. Breach duration
3. Cooldown period (scaling out and scaling in)
4. Time periods
For example, if you know that your application is heavily used on Christmas Day, you can add a specific start and end date along with the number of instances you want your application to scale out to.
Review the policy configurations, and make sure to enable the policy.
Perform the load test using Load Impact
Navigate to the Load Impact service, and open the Load Impact dashboard. Create a test and add the link to the application, as shown in the following screen capture:
Add a scenario, and run the test to successfully test it out, as shown in the following screen captures:
You can review the scaling history in the Auto Scaling service for the implications of the Load