4 More Underrated Kubernetes Tools and Features in AWS for DevOps Engineers
Explore four lesser-known Amazon EKS tools and features for Kubernetes cost optimisation, intelligent node scaling, network security, and AI-assisted troubleshooting.
Series: Underrated Kubernetes Tools (2 articles)
- 24 More Underrated Kubernetes Tools and Features in AWS for DevOps Engineers This article
Running Kubernetes workloads on AWS requires more than deploying containers and maintaining cluster availability. Production environments also need cost visibility, efficient infrastructure scaling, network security, and effective troubleshooting.
Amazon Elastic Kubernetes Service (Amazon EKS) provides several capabilities that help teams address these challenges. Some are familiar to experienced engineers, while others are underused or overlooked in day-to-day operations.
In this article, we explore four useful Kubernetes tools and features in AWS that can help DevOps engineers improve operational efficiency and manage production workloads more effectively.
1. EKS Split Cost Allocation – Understand Where Kubernetes Money Goes
A Kubernetes cluster can look healthy while quietly wasting money.
For example, an application may request excessive CPU and memory, or several teams may share the same cluster without a clear understanding of their infrastructure costs.
AWS provides split cost allocation data for EKS, allowing eligible Kubernetes-related costs to be analysed through AWS billing tools. Kubecost is another option for more granular Kubernetes cost visibility.
Useful questions include
- Which workloads account for the largest share of costs?
- Are teams requesting more CPU and memory than they need?
- How much infrastructure spending is associated with individual applications?
- Which workloads might benefit from rightsizing?
Why is it useful?
- Improves cost transparency.
- Supports internal chargeback and showback.
- Helps identify optimisation opportunities.
Cost allocation requires appropriate configuration, and not every charge will necessarily map neatly to an individual pod.
2. Karpenter – Provision Nodes Based on Workload Requirements
Kubernetes scheduling and infrastructure scaling are related, but they are not the same thing.
The Kubernetes scheduler decides where pods should run on available nodes. A node provisioning system must provide additional suitable capacity when existing nodes cannot accommodate workloads.
Karpenter helps solve this problem by provisioning and consolidating compute capacity based on workload requirements and configured constraints.
For example, a workload may require additional memory, a particular CPU architecture, or a specific class of instance. Karpenter can help select suitable EC2 capacity within the configured constraints.
Why is it useful?
- Responds to changing workload requirements.
- Can reduce idle compute capacity through consolidation.
- Supports flexible EC2 instance selection.
Karpenter is especially useful for variable workloads, but consolidation policies, disruption budgets, scheduling requirements, and availability needs must be configured carefully.
3. Kubernetes Network Policies – Control Which Pods Can Communicate
Network security is often overlooked when teams focus primarily on IAM roles, security groups, and public endpoints.
Kubernetes NetworkPolicies provide a way to define which pod-to-pod or pod-to-external communication is allowed. However, policies only take effect when the cluster's networking implementation supports and enforces them.
On EKS, the choice of networking implementation matters. AWS documents supported networking options and their limitations across different EKS compute environments.
For example, a team may want to restrict a payment service so that only approved application pods can connect to it.
A NetworkPolicy can express that intent, provided the installed networking implementation supports the required policy behaviour.
Why is it useful?
- Helps isolate sensitive workloads.
- Reduces unnecessary internal network access.
- Supports defence in depth for microservices.
Do not assume that creating a NetworkPolicy automatically enforces traffic restrictions. Verify support in your networking configuration and test the resulting behaviour.
4. Amazon Q for EKS – Investigate Cluster Problems Faster
Troubleshooting Kubernetes can involve multiple layers, including application configuration, pod scheduling, node health, IAM permissions, networking, and AWS infrastructure.
Amazon Q Developer integration in the EKS console can assist with investigating supported cluster issues by analysing errors and warnings and suggesting potential mitigation steps.
This is particularly useful when an error message does not clearly identify its root cause.
For example, if a workload cannot start, an engineer can begin with standard Kubernetes commands
1
2
3
4
kubectl get pods -A
kubectl get events -A
kubectl describe pod <pod-name>
kubectl logs <pod-name>
These commands provide direct evidence about pod status, scheduling events, configuration errors, and application logs. AI-assisted investigation can then help interpret the findings and suggest additional troubleshooting steps.
Why is it useful?
- Helps narrow down possible causes.
- Reduces time spent searching documentation.
- Provides a starting point for investigating unfamiliar EKS errors.
AI-assisted troubleshooting should complement, not replace, log inspection, testing, and engineering judgement.
So....
Effective Kubernetes operations on AWS require engineers to balance infrastructure costs, workload scaling, network security, and troubleshooting efficiency.
EKS split cost allocation can improve cost visibility, Karpenter can help provision compute capacity dynamically, NetworkPolicies can strengthen workload isolation, and Amazon Q Developer can assist with supported troubleshooting workflows.
The best approach is not to adopt every available tool. Instead, identify the most significant operational challenges in your environment and select tools that address them.
By combining these capabilities with sound Kubernetes practices, DevOps teams can improve the reliability, security, and cost efficiency of their AWS environments.
Series: Underrated Kubernetes Tools (2 articles)
- 24 More Underrated Kubernetes Tools and Features in AWS for DevOps Engineers This article
Enjoyed reading this content? Let the author know!
Your likes, comments, shares, and saves help creators reach more builders.
Loading recommendations
Loading article