We are thrilled to announce that Turbonomic now supports Google Cloud with the latest release of 8.4.2. This is a huge step in supporting our customers who want to take advantage of the unique services that the Google Cloud Platform (GCP) has to offer.
For years now, Turbonomic has provided support for customers leveraging AWS and Azure, helping them assure continuous application performance at the lowest possible cost.
More and more customers are taking a multicloud approach, with Google Cloud playing a key part in their business transformation strategies. Today, if you’re leveraging one, two, or all three of these major cloud providers, Turbonomic’s continuous cloud optimization can help. Let’s dive into what exactly this means for you.
GCP Continuous Cloud Optimization
Turbonomic takes a performance-first approach in its analysis: it will never sacrifice the performance of a business application simply to cut costs. By giving applications exactly what they need to perform, customers realize significant short (and long-term) savings as they drive greater elasticity across their cloud estate.
Turbonomic is about taking existing data and turning that into actions you can execute and automate with confidence. Let’s start with what data we’re pulling in from GCP.
Turbonomic discovers GCP projects that define compute, storage, and networking resources for your workloads. If the service account has organizational level permissions, Turbonomic can discover your complete resource hierarchy allowing our software to provide a more in-depth, full-stack analysis of your GCP environment.
Once the GCP target has been enabled and configured, Turbonomic will automatically identify resources utilized by each Virtual Machine from Virtual Memory (vMem) to Virtual CPU (vCPU) with respect to vCPU capacity utilization. Turbonomic also automatically identifies the attached storage, including storage utilization, IOPS, and throughput.
By giving the service account Billing Account Viewer permissions (instructions found in GCP 8.4.3 Help Center), you'll have visibility into your GCP bill and associated service costs over time. This includes negotiated discounts, and GCP's own committed use discounts - very similar to AWS and Azure's reservations/savings plans capabilities.
Like any public cloud provider, the underlying host infrastructure doesn’t natively report memory metrics of the virtual machines. An agent is required so they can be accurately collected, in this case, the GCP Ops Agent needs to be installed on each Virtual Machine to collect memory metrics.
Elasticity starts with actions. What types of actions are available?
Again, it’s all about turning data into actions. Turbonomic analyzes the VM resourcing, as well as billing data, to generate specific trustworthy actions. The two types of actions available today are Scale VM and Reconfigure VM. (In future releases, Turbonomic will be able to make scaling and performance recommendations based on the “committed use discount” for your GCP compute engine workloads.)
Scale VM Action
For Scale VM, Turbonomic will recommend scaling VMs to optimize performance for the business application while simultaneously taking cost into consideration. Utilization percentiles, workload costs, and scaling constraints are also analyzed to recommend accurate scaling actions.
Watch this quick walk-through of GCP visibility & actions.
Reconfigure VM Action
As for the Reconfigure VM action, GCP provides a specific set of machine types for each zone in a region. If you create a policy that restricts a VM to certain machine types and the zone it’s currently on does not support all those machine types, Turbonomic will recommend a reconfigure action to notify you of the non-compliant VM. For example, assume Zone A does not support machine types for the M1 family. When a VM in that zone applies a policy that restricts it to M1, Turbonomic will recommend that you reconfigure the VM.
Got GKE? We’ve supported Google Kubernetes Engine for some years now, but there are several advantages to “stitching” your GKE clusters to the underlying GCP resources.
Understand the costs associated with your GKE clusters. This is especially useful for organizations who are making a strategic investment in the platform. Those clusters will horizontally scale to support more and more applications. If the clusters are multi-tenant, this cost visibility can be used to generate reports for showback. Let’s digitally transform responsibly!
Confidently suspend unneeded nodes. Turbonomic’s cluster compute analysis will determine when you can suspend unneeded nodes. You can have confidence in the analysis because it’s not based utilization, but rather takes the approach of proving you can run your applications on the remaining nodes (this applies to the real time analysis as well as plans). By stitching GKE to GCP you see the cost savings and by using only the cloud resources you truly need you’re supporting environmentally sustainable IT.
Scale responsibly (and support environmental sustainability). Likewise, when Turbonomic drives actions to horizontally scale nodes, you can see the associated GCP costs to understand the investment required, both in real time and in plans.
As excited as we are about this first release, there’s lots more coming. In future releases, our customers will have the ability to simulate GCP workload migrations, and cloud optimization plans. Many of our customers as well as early results from our 2022 State of Multicloud Survey (the report will be published in the next month or so, you can check out the 2021 report here) tell us that migrating more workloads to public cloud is the most important initiative they are currently tackling. Additionally, we plan to build out support for folder hierarchy browsing and scoping, cost by tag reporting, storage delete actions, and more! And, as noted earlier, Turbonomic will be able to generate scaling and purchasing actions based on your committed use discounts.
For more details on Turbonomic’s 8.4.2 release and what technologies we support, check out our “Turbonomic 8.4.2 Target Configuration Guide”.
How to Get Started
The first step is to add GCP as a target, the GCP Probe would need to be enabled for your Turbonomic Instance, if you have SSH access to the Turbonomic Instance or Virtual Machine.
If you don’t know how to enable the GCP Probe, check out our tutorial, Enable GCP Probe with Turbonomic, on how to do so. And if you’re using Turbonomic’s SaaS offering, please contact your account team who can enable GCP for you.