Blog

Magalix Node Advisor for Kubernetes is now in Public Preview

Magalix Node Advisor allows Kubernetes users to right-size their clusters, for the best application performance and lowest cost.

 

Magalix Node Advisor 2.0 is an advanced analytics service to optimize your Kubernetes cluster. It saves you hundreds of hours figuring out the right combination of nodes to achieve the best performance, highest utilization, and lowest cost. You’ll get the best combination of node types and sizes. You won’t need any specialized knowledge to understand the performance and cost implications of these recommendations. You’ll also get in advance the analysis of expected utilization based on past and future workloads. No complex calculations to figure out the best billing plan and how much you can save. Magalix automatically analyzes your cluster workloads to match it against best capacity and performance.

Register now and get 14-days of free Node Advisor reports. 

What Problem Does the Node Advisor Solve?

It is a tedious, time-consuming task to figure out the best type, size, and combination of Kubernetes nodes. You have to look at many performance and infrastructure utilization metrics to estimate the amount of capacity you need. It also requires you to have a deep understanding of different types of provider virtual machines (VMs), their performance, and cost implications. For example, AWS has more than 100 instances types and sizes and the list gets longer every few weeks. It becomes tricky to calculate utilization and resulting buffer inside your capacity with the new nodes and capacity.

How Does The Node Advisor Elegantly Solve the Capacity Issues of Kubernetes Clusters?

Magalix provides a detailed analysis of cluster capacity and shows cost and utilization improvements you could achieve.

Magalix Node Advisor analyzes your current capacity and dozens of metrics to characterize workloads. It figures out memory and CPU patterns to decide on capacity changes in your cluster. Magalix Node Advisor may recommend the need for more cores in fewer VMs but less overall memory for example. The proposed changes are set to achieve three main goals:

  1. Improve cluster and node utilization and tweak it to your targets.
  2. Improve containerized application performance whenever needed by having enough capacity to avoid CPU throttling or Out of Memory (OOM) crashes
  3. Reduce the overall cost of infrastructure by either selecting better VM families or reducing needed capacity.

 

 

Get the Best Combination of Nodes and Capacity

Magalix Node Advisor gives you detailed recommendations of the best node combinations you should provision to save money.

To realize the best utilization and lowest cost, you get a detailed plan to create node pools, VM types and VM sizes. Even if you just have the right capacity, the Node Advisor can still save you a lot of money. It tweaks VM types and shuffles capacity within your cluster. For example, in the case of memory intensive containers, you will save 40% if you change your nodes from general purpose instances (M5) to memory intensive instances (R4).

Compare Different Billing Plans

You can now compare different billing plans of your current provider to other cloud providers and select the best that works for your needs. You can now see the updated pricing of your proposed capacity. For example, if you run Kubernetes on AWS, the Node Advisor compares the pricing of the on-demand, reserved instances, and spot instances.

“A $1.5 hourly savings in a 20 nodes cluster saves our users up to $14,000 yearly.” Mohamed Ahmed, CEO - Magalix

 

 

You Are in Full Control of Optimization Policies

Tweak your preferences and update the Node Advisor optimization policies. You can set restrictions on the number of nodes and the buffer size you want inside your cluster. For example, if you set the buffer size to be 30%, Magalix Node Advisor will recommend infrastructure with 70% utilization. You can tweak the policy to optimize for maximum reliability, i.e distribute containers on as many nodes as possible. Or tweak it to increase the density of containers and reduce the number of nodes.

“I was skeptical about Magalix Node Advisor at the beginning. But I was surprised by its accuracy and potential savings that we can achieve.” Amr Farid, DevOps engineer.

Node Advisor Supports All Major Public Cloud Providers

Magalix Node advisor supports AWS, Azure, GCP, and IBM. It is always up to date with the latest VM families - see detailed supported lists here. You will always be on top of your anticipated cost since Node Advisor continuously updates the pricing.

 

 

Optimize Your Kubernetes Capacity Today

Register and connect your existing Kubernetes cluster with a single kubectl command. Within a few minutes, cluster metrics will start flowing and our AI engine will engage shortly analyzing your containers and infrastructure. You will get 14-days free full access trial.

 

Frequently Asked Questions (FAQ)

Do you have any plans to automate the Node Advisor recommendations?

Yes. It is a quite complex process to upgrade the capacity of your cluster. We working on some basic scenarios that will help you reach maximum efficiency with full automation. Stay tuned!

How frequently do you update the pricing of different cloud providers?

We update the pricing and families types as the cloud providers make changes to their billing or pricing model.

Do I need to provide Magalix access to my cloud provider account in order for the Node Advisor to work properly?

Not at this point. We use the existing standard pricing to estimate the cost of the current recommended capacity. This does not require any special access to your cloud provider accounts.

Does the Node Advisor consider pricing differences between different regions?

Yes. It assumes that you will keep different your cluster at the same time zone.

Any plans to provide analysis across different providers?

Not yet. We would love to hear from you. Please share your thoughts here.

Does the Node Advisor Support clusters in different regions?

Not at the moment. It will pick a region and optimize cluster capacity and pricing.

Does the Node Advisor Support Managed Kubernetes?

Yes. It supports Google Kubernetes Engine (GKE), AWS Elastic Kubernetes Service (EKS), and Azure Kubernetes Service (AKS)

 

Mohamed Ahmed

Jun 11, 2019