A big challenge with cloud services is keeping costs under control, and business expects IT Infrastructure and Operations to be doing just that, rather than focusing solely on engineering and operations
Analysts Explore Top Industry Trends 2018 - Gartner
Why keeping costs down is difficult
Each department in the company may use its own cloud infrastructure
Cloud resources are not adjusted to the company’s current needs
Cloud spend is not monitored
Long-term cost management strategy is lacking
Cloud resource peak capacity is rarely used
Our approach to cloud cost management
Organize
1
Group all resources into organization
Prepare and introduce a resource structure
Analyze
2
Summarize pain points
Research corrective actions
Improve
3
Billing Management
The use of resources and services
Why us
We have experience managing and optimizing multi-cloud environments
Working for industry leaders
Providing teams of top-notch cloud experts that solve problems efficiently
Delivering projects on time and on budget
We’re on top of advances in cloud computing
Providing technology advice and support
Case study: cloud cost optimization for Kubernetes cluster
1
Case
Multiple GKE clusters created for development purposes
3-6 nodes, 1vCPU, 3GB RAM
Clusters are expected to be short-lived and redeployable at any time
2
Problem
Clusters stay online for extended periods of time
Nodes cost about $25 per month each while not reaching peak capacity
3
Solution
Introduce cleanup system, unless cluster is marked to stay
Use preemptible VMs with node pool autoscaler (~$5/month each)
Consolidate projects to qualify for sustained use discounts (730 hrs/month, stacked from multiple VMs) on Google Cloud Platform - surprisingly, it is sometimes cheaper not to turn machines on and off
Long-distance improvements: committed use discount