In today's cloud-driven world, managing costs is not something you do every now and then, but it has become rather crucial for a business to understand where to money goes and how to handle these costs. Companies are setting up FinOps teams, CCoE’s, having engineers track costs and the list goes on.
That's why we at IBM Software believe Turbonomic helps you successfully manage your cloud spend. We're pleased to announce a new feature that optimizes AWS Redshift costs using Turbonomic's SUSPEND actions.
This blog will help you understand what AWS Redshift is, how you can optimize it using Turbonomic and who this is for.
This new feature, released in H2 2024, provides intelligent automation to help you reduce your Redshift costs by identifying and suspending idle clusters.
Understanding AWS Redshift:
Before diving into the details of this new feature, let's briefly recap what Amazon Redshift is. It's a fully managed, petabyte-scale data warehouse service in the cloud, designed to analyze vast amounts of data using your existing business intelligence tools. Whether you opt for provisioned clusters or the serverless option, Redshift provides the power and scalability you need for your analytics workloads.
That was the tricky part...
And if you’re going to explain this to someone in easy terms, well here we go; Redshift is like a super-powered spreadsheet in the cloud that keeps track of everything about your sales. It helps you quickly see what's selling best, who your best customers are, and where you can improve, so you can make more money. I hope that helped. Now let’s dig deeper...
Provisioned vs. Serverless:
Provisioned clusters offer predictable performance for consistent workloads.
Redshift Serverless provides on-demand scaling, eliminating the need to manage infrastructure.
Provisioned clusters are like a reserved restaurant table, offering consistent performance for predictable, steady workloads. Redshift Serverless is like a food truck, automatically scaling and handling infrastructure so you only pay for what you use, ideal for fluctuating needs. Easier to understand I guess if we put it that way.
At Turbonomic we focus on different personas, such as Cloud / DevOps Engineers. Let’s assume you’re one of these “Personas” and you’re reading my blog. Understanding these differences between provisioned and serverless is crucial for aligning Redshift deployment with application needs and cost optimization strategies.
Introducing Turbonomic's Redshift Suspend Actions:
The newly introduced Suspend Actions feature provides a powerful tool to eliminate unnecessary costs by identifying idle clusters. Turbonomic discovers Redshift provisioned clusters through your AWS targets, and represents them as data warehouse entities in the supply chain.
Imagine Redshift is like a car you rent. If you're not using it, you're still paying for it. This new feature, "Suspend Actions," helps you turn off those "rental cars" (Redshift clusters) when you're not using them, saving you money. Turbonomic is like a smart manager that looks at your AWS account and sees all your Redshift "cars." It shows you which ones are sitting idle. Sounds good right? Below in the screenshot we can read in the note that Turbonomic does not discover Redshift Serverless. This is important to keep in mind.
So how does it work? Well, Turbonomic will generate a suspend action for a Redshift cluster if the following conditions are met:
The cluster's clusterState and clusterAvailabilityStatus are both "Available" (meaning the cluster is powered on and running). Suspend actions are generated for single-AZ deployments. What about multi-AZ? For multi-AZ Redshift entities, Turbonomic can generate suspend actions only after the resource has been updated to a single-AZ deployment type. This is a mandatory prerequisite.
On top of that Turbonomic will generate a suspend action for a Redshift cluster if the cluster has been idle for a defined amount of time. More to that later in this article.
Now let’s come back to the “Personas” we’re having in mind while writing this article. What are some key benefits for Cloud / DevOps Engineers? Let me list them for you in the following order:
Cost Savings: By automatically suspending idle clusters, you can significantly reduce your Redshift spending, a key metric for demonstrating value to stakeholders. Very important metric indeed.
Automation: Turbonomic's intelligent automation eliminates the need for manual monitoring and intervention, freeing up your time for more strategic tasks. This aligns with DevOps practices of automation. In other words, consolidate using data, automate using tools and motivate with leaderboards.
Policy-Driven: Turbonomic ships with default automation policies, and you can create custom policies to tailor the feature to your specific needs, giving you granular control.
Tracked Savings: Turbonomic tracks the savings achieved through these actions, providing clear visibility into your cost optimization efforts, aiding in reporting and demonstrating ROI.
Visibility: You gain visibility into the entire application stack related to Redshift, helping in troubleshooting and performance optimization.
By now you probably have some questions about permissions and what idleness means. Let me explain this to you as follows:
"Idleness" refers to the time a Redshift cluster has had no active database connections. The default is 24 hours, but this can be adjusted in your policies. You can tailor this to your specific workload patterns.
To leverage this feature, ensure your Turbonomic IAM role or user has the following permissions:
redshift:DescribeClusters
Suspend actions are available starting with Turbonomic version 8.14.5 and above.
Turbonomic's new Redshift Suspend Actions feature empowers you to take control of your cloud costs by automating the suspension of idle Redshift clusters. Think of Turbonomic as a valet who notices your Redshift "car" (cluster) sitting idle. With the new Suspend Actions, they automatically park it, saving you rental fees until you need it again. Basically, you get the ride when you want, without paying for the parking. No need for Uber, just Turbo!
Upgrade to Turbonomic 8.14.5 or above and start optimizing your AWS Redshift costs today!
If you have any questions or need assistance, please don't hesitate to reach out to Samuel Nord.