From staff capacity to cash flow, managing your resources is critical to keeping your business running smoothly. And the same applies to your Cloud use.
As the keeper of your data and critical infrastructure, your Cloud must be able to respond to the ebbs and flows of your usage — expanding when you need more and pairing down when you don’t.
This is why Azure created its Autoscaling feature — to give you the flexibility in performance and cost that you need in the Cloud.
Let’s take a look at how it delivers and go over a few things to help you get the most out of the feature.
Azure autoscaling: only pay for what you need
Imagine if every Friday you needed to order donuts for your weekly bocce ball club meeting. The number of donuts you need fluctuates week to week depending on member availability. Sometimes 15 or more people show up, sometimes just a handful.
Now imagine if your baker only made it possible for you to buy one dozen donuts each week. On those low attendance weeks, you’d find yourself swimming in sugary excess. And on the high end, dealing with some major sugar hangry cases.
The simple solution, of course, would be for your baker to flexibly adjust the order to meet your needs.
And the same should apply to your Cloud solution.
Azure Autoscaling does just that by automatically scaling up and down, depending on your needs. By doing so, Azure ensures you aren’t spending money where resources aren’t needed.
Let’s say, for example, your Cloud application is busiest from 6 am to midnight and slows dramatically between midnight and 5 am. In this scenario, it would be foolish to have all your virtual machines (VMs) running at full speed over the 24-hour period — you need scaling that matches your needs.
And responsive scaling is what Azure’s Autoscaling excels at, allowing you to configure Windows Azure to automatically scale your application without manual manipulation.
By doing so, Azure ensures your ideal performance is maintained while also only charging you for what you use.
Now, let’s look at a few ways Azure Autoscaling helps you control those costs…
Understanding the metrics
Metrics are an important part of your Azure usage and understanding how they work will help you keep your costs down.
Azure Autoscaling operates on two types of metrics: resource metrics and custom metrics.
Here’s a high-level explanation of what they mean:
1. Resource metrics
Resource Metrics apply to anything related to your usage within Azure. This includes memory usage, CPU usage, disk usage, thread counts, and queue length.
Using these metrics, you can set Azure Autoscaling to scale up or down based on your usage parameters. Doing so will ensure you’re only paying for the usage you need and will keep your costs down as a result.
2. Now let’s look at custom metrics
Unlike resource metrics, which look at usage within Azure, custom metrics are external metrics pointed into Azure from your application(s).
An example of a custom metric would be enabling Azure’s time feature to check the number of inbound and outbound API instances to and from your application. Azure will then analyze the minimum and the maximum number of instances you’ve preset. If the instances go above your maximum, it will add extra units to accommodate the workload. The next day, however, Azure will reset to your parameters to ensure you’re not paying for unused units.
Like resource metrics, properly setting your custom metrics with Azure Autoscaling is an important component in flexibly scaling with your needs and keeping your costs down as a result.
Horizontal vs. vertical scaling
Another key concept in controlling your Cloud costs is understanding the difference between horizontal and vertical scaling in the Cloud.
Scalability in the Cloud means being able to expand or shrink the number of machines being used to match current needs.
What’s the difference between horizontal and vertical scaling you ask? Let’s take a look.
Horizontal scaling:
This sort of scaling increases or decreases the machines within your Cloud resource pool — switching machines on, when you need them and switching them off, when you don’t.
Virtual scaling:
By comparison, keeps the same number of machines but increases or decreases the power of those machines based on your needs.
With virtual scaling, the same number of machines is always running — meaning you’re paying for them even when you don’t need their full capacity.
And here’s where the favorability of horizontal scaling comes to light. By automatically turning unneeded machines off, horizontal scaling ensures you’re only using what you need and thereby only paying for what you need as a result…and did we mention Azure Autoscaling uses a horizontal scaling model?
Start saving money today with Azure Autoscaling. Contact us now to get started.
