Kubernetes poses several cost-related challenges. Fortunately, teams are working hard to help engineers take full advantage of Kubernetes without making their CFOs faint at the sight of their cloud bills.
CAST AI and Harness are two examples of such tools. But which one is a better pick for teams looking to reduce their cloud expenses for running Kubernetes? Keep on reading to find out.
CAST AI – cloud native cost optimization platform
Harness – automated cloud cost management
Created by cybersecurity experts, CAST AI is an ISO 27001-certified, comprehensive cloud automation platform for optimizing Kubernetes environments. Companies across e-commerce and adtech are using CAST AI to cut their cloud bills by 50% to even 90%.
Harness is a continuous integration and delivery platform for engineering and DevOps teams. The platform includes a cloud cost management module that supports cost savings in Kubernetes and offers Business Intelligence tools for analyzing cloud costs.
CAST AI vs. Harness – quick feature comparison
Detailed feature comparison of Harness and CAST AI
- Cost visibility
- Cost optimization and automation
- Spot instance automation
- Full multi cloud optimization
1. Cost visibility
Cost allocation and reporting
In CAST AI, cloud costs are split into project, cluster, namespace, and deployment levels. Teams may track expenses down to individual microservices and then produce a full estimate of their cluster costs. The platform uses industry-standard metrics that work with any cloud service provider.
CAST AI‘s cost allocation works on a per cluster and per node basis. The team intends to expand the cost dimensions disclosed to include control plane, network, egress, storage, and others. In addition, a capability for ongoing cloud cost reporting is being developed.
Harness offers in-depth Kubernetes visibility by displaying the utilized, idle and unallocated resources per workload and cluster. This data is paired with ready-made insights to help teams make the right decisions when taking action. Users can correlate costs generated by deployments and change the in replica count, CPU, or memory configurations, with the help of cost events and specific lines of code.
Moreover, Harness visualizes cost information by projects, teams, business units, departments, and more. Users can also get periodic reports on cost and usage metrics that matter to them most. Moreover, governing usage is easier thanks to custom budgeting, spend forecasts, and accounts for cost showbacks and chargebacks.
The CAST AI team is now working on real-time warning capabilities that will tell customers when their cloud spending exceeds a certain level, reducing the danger of an out-of-control service bill.
Harness identifies anomalies in cloud usage and informs users every time their expenses are greater or lower than they should be. The platform achieves that by analyzing historical spend trends.
Cost view across multi cloud
Multi cloud support is a key part of cost optimization efforts since many companies nowadays use more than one cloud platform.
Allocating expenditures for multi-cloud setups is tricky, but CAST AI‘s enhanced multi cloud capabilities make it much easier. Thanks to universal metrics from Grafana and Kibana, the platform easily operates with any cloud service provider and enables cross-cloud visibility.
Harness allows users to keep track of their cost and usage across all of their cloud resources. The tool displays insights on custom, data-driven dashboards to help teams track resources across every cloud provider they use.
2. Cost optimization and automation
CAST AI – fully automated cost optimization
Rightsizing with automated instance selection
CAST AI chooses the best instance types and sizes to meet an application’s requirements while reducing cloud costs. When a cluster needs extra nodes, the automation engine selects the instances with the highest performance at the lowest cost. Given that everything is automated, teams don’t have to do anything extra.
The platform enables multi-shape cluster formation as using the same instance shape for every node in a cluster can easily lead to overprovisioning. CAST AI offers the right mix of multiple instance types for the application’s requirements.
Horizontal pod autoscaling and node autoscaling
CAST AI automates pod scaling settings to help companies avoid overprovisioning their infrastructures.
The Horizontal Pod Autoscaler calculates the optimal number of necessary pod instances based on business KPIs. If there is no work to be done, the feature gradually reduces the replica count of your pods until it reaches 0 and then removes all pods.
CAST AI also ensures that the number of nodes in use always matches the application’s needs, scaling nodes up and down automatically.
Cluster scheduling and termination
CAST AI automatically pauses and resumes clusters formed inside the platform so that teams avoid paying for resources they don’t use.
Bin packing is done automatically.
Kubernetes presents a financial challenge since it distributes apps evenly throughout a cluster, regardless of how cost-effective this architecture may be.
CAST AI alters the default pod scheduling strategy and employs automated bin packing to maximize savings based on user preferences. The cost reductions are larger when there are fewer nodes.
Harness – recommendations combined with automation
Detailed reports and recommendations
Harness provides its users with in-depth reporting combined with practical suggestions for cluster utilization, rightsizing, autoscaling, cleaning underutilized or orphaned resources, and picking the optimal cloud service tier based on usage trends. Before implementing the suggestions, users can perform a what-if analysis.
Another Harness feature that helps to optimize costs is the option to set monthly, quarterly, and yearly budgets or keep track of expense variations. Thanks to precise forecasting, teams can see whether their present spend rate is within budget or on its way to surpassing it. The alerts sent at key thresholds offer some extra help in tracking the budget consumption.
Harness offers a range of automated features as well.
Automation in Harness:
- Since non-production resources are usually used during working hours, teams can use the AutoStopping feature to turn them off anytime they’re not in use. If workloads are running on Spot instances, there’s no risk of disruption if users execute them on coordinated spot instances.
- Harness Continuous Delivery provides cost information about your apps, services, and environments without the need for human tagging. This cuts a lot of time and effort teams would dedicate to this task to improve cost allocation and reporting.
3. Spot instance automation
Spot instances offer massive cost savings when compared to the cost of On-Demand instances – even up to 90%. But providers may reclaim these resources back at any time. That’s why automation is so important for teams that want to take advantage of Spot instances.
In CAST AI, the replacement of interrupted spot instances is entirely automatic. Teams no longer have to worry about their apps running out of capacity. To achieve high availability, the platform constantly looks for the best instance alternatives and spins new instances in a fraction of a second.
Harness recognized the value of Spot instances as well. Its users can run workloads on fully orchestrated Spot instances without having to worry about the interruption. The platform handles this part smoothly.
4. Full multi cloud optimization
The need to analyze, manage, and optimize cloud costs across providers is more important than ever as we enter the era of multi cloud.
CAST AI satisfies this requirement with a variety of multi cloud features:
- Active-Active Multi Cloud – the platform distributes apps and replicates data over several cloud services so that if one fails, others keep the applications running and business continuity is guaranteed.
- Global Server Load Balancing – CAST AI intelligently distributes traffic across all clouds in use, always selecting up and healthy endpoints.
- Multi cloud visibility – the solution delivers cost allocation across cloud services thanks to data from Grafana and Kibana.
Harness currently only offers cost visibility for AWS, Microsoft Azure, and Google Cloud.
Users of CAST AI can begin by using the free Cluster Analyzer to see whether they can save money on their cloud services. The read-only agent evaluates their infrastructure and shares helpful recommendations. Next, you can either apply these insights manually or use automated cost optimization features. In the latter scenario, users can choose between two plans: Growth and Enterprise. CAST AI helps to achieve cost savings of at least 50%.
Harness offers three plans: Free, Team, and Enterprise. Note that the first two aren’t publicly available yet. In the Team plan, Harness charges 2.25% of the customer’s annual cloud spend. In the Enterprise plan, the company charges 2.50% of yearly cloud expenses.
Summary: CAST AI vs. Harness
Both Harness and CAST AI are excellent cloud cost management platforms that improve cost visibility, allocation, monitoring, and optimization.
While Harness offers in-depth recommendations and a limited number of automation features, CAST AI provides teams with a number of handy automation features that guarantee cost savings and streamline work with Kubernetes.
The comprehensive automation features, unique multi cloud functionality, and cloud native architecture position CAST AI as the top cloud cost optimization platform.
P.S. If you prefer a hands-on approach, you can always run the free CAST AI Cost Analyzer to see what the platform could save you automatically.