Turbonomic Parking: Where intelligent automation meets limitless potential. Prepare to unlock potential you didn't even know you had with Turbonomic Parking now...
Turbonomic Parking extended its capability for Public Cloud (AWS, Azure & Google) Scale group services!
Now you can Park,
- AWS Auto Scaling Groups (ASGs who are independently deployed) -- Available from 8.13.1
- Azure Scale Sets -- Available from 8.13.6
- Google Managed Instance Groups (MIGs) -- Available from 8.17.3
Note: If you're in Turbonomic 8.17.3, Yay! You get all the three features!!
How Scale Groups are used in the Organizations?
The Scale group service family belong to the IaaS offering from the Cloud Providers, who are the alternatives to the Serverless & managed PaaS offerings (like AWS Lambda, Azure Functions & Google Cloud Run). They are considered over the Serverless services, due to the perks of high control over large scale predictable workloads and heavily customized environments
Where is the Opportunity for Cost Savings?
The primary cost optimization opportunity stems from implementing a methodical parking strategy, which entails the scheduled, graceful shutdown and subsequent restart of compute assets. Aligning resource availability with actual demand in this manner eliminates expenditures on idle capacity. For non-production workloads, this technique can reliably deliver cost reductions ranging from 40% to 60%.
Use Case for Cloud Cost Savings -- Running Compute Scale Groups (AWS, Azure & Google)?
Auto-scaling groups like AWS ASGs, Azure Scale Sets, and GCP MIGs are built to adapt—but during idle hours, they often hold onto excess capacity, quietly inflating your cloud spend.
Applications running on Azure Scale Sets experiences high traffic during business hours but drops significantly overnight
Cloud Providers native solutions for Parking & Challenges faced by the Organizations?
Manual Complexity & Limited Automation
Native solutions often depend on custom scripts and manual intervention to stop instances. This adds operational complexity, especially when trying to identify idle resources or align actions with business operating hours.
Scalability & Change Management Challenges
Manual processes don’t scale well across large environments. They introduce friction in change management, making it difficult to implement consistent cost-saving strategies across dynamic workloads.
Lack of Fine-Grained Control
There’s minimal native support for selectively stopping specific instances based on real-time utilization or business context. This limits the precision of optimization efforts and leaves potential savings on the table.
Real World Usage & Value Proposition
Scenario: Distributed and Stacked Business Application & Services running across different VMs in Scale Groups, those needs Business hour awareness, Granular control towards Idle Instance detection as a continuous process
Goal: Identify idle duration, business hours & execute policy driven parking schedules to automate parking actions at scale
Value: Realize cost $ savings over the Parked Compute resources within every Scale group
How Scale Group services appear in Turbonomic parking UI:
Azure Scale Set: Member Instances view
Google MIG: Member Instances view
Turbonomic exclude Scale Group workloads that belong to Cloud Kubernetes Services, from Parking. Why?
Scale Groups can be a part of (say) AWS EKS, Azure AKS & Google GKE, as they are a foundational component of how Kubernetes manages its Compute resources.
- Kubernetes Cluster attempts to scale their Node Pools (Scale Groups), for various reasons like Pods scheduling, Node provisioning...
- Pod distribution and operational considerations needs to be evaluated before providing the option to park individual Nodepool (say)
Hence with the latest Turbonomic Version, Parking does not show the scale groups those are a part of Kubernetes services (EKS, AKS, GKE). However, there would be a future implementation plan that would bring additional value of Parking scale groups in Kubernetes clusters.
Ready to Go?
Try out by upgrading your Turbonomic Version to 8.17.3 & Above and begin your continuous cost optimization journey with Parking.
For new users, please reach out to our IBM representative and try a Demo!
Additional resources: (Parkable workloads - virtual machine specs)
What's Next
For more feedback/Add Ideas..