KubeBuddy

0.0.24

Private/aks/checks/ResourceManagementChecks.ps1

                                $resourceManagementChecks = @(

    @{

        ID          = "AKSRES001";

        Category    = "Resource Management";

        Name        = "Cluster Autoscaler";

        Value       = { $clusterInfo.properties.autoScalerProfile -ne $null };

        Expected    = $true;

        FailMessage = "Cluster autoscaler is disabled, requiring manual node scaling that leads to over-provisioning during low demand (wasted costs) or under-provisioning during peak times (performance issues and pod scheduling failures). This significantly increases operational overhead and cloud spending.";

        Severity    = "Medium";

        Recommendation = "Enable Cluster Autoscaler using 'az aks update --resource-group <rg> --name <cluster> --enable-cluster-autoscaler --min-count <min> --max-count <max>' on node pools. Configure appropriate min/max node counts, scale-down parameters, and node pool priorities for optimal cost and performance balance.";

        URL         = "https://learn.microsoft.com/azure/aks/cluster-autoscaler";

    },

    @{

        ID             = "AKSRES002";

        Category       = "Resource Management";

        Name           = "AKS Built-in Cost Tooling Enabled";

        Value          = { $clusterInfo.properties.metricsProfile.costAnalysis.enabled };

        Expected       = $true;

        FailMessage    = "Cost analysis and OpenCost integration is disabled, providing no visibility into per-namespace, per-workload, or per-application spending. This makes it impossible to implement cost allocation, identify expensive workloads, optimize resource usage, or implement chargeback policies for different teams.";

        Severity       = "Medium";

        Recommendation = "Enable cost analysis using 'az aks update --resource-group <rg> --name <cluster> --enable-cost-analysis' to track namespace and workload-level costs. Use the cost insights to identify expensive workloads, optimize resource requests, and implement chargeback/showback policies.";

        URL            = "https://learn.microsoft.com/azure/aks/cost-analysis";

    },

    @{

        ID             = "AKSRES003";

        Category       = "Resource Management";

        Name           = "Vertical Pod Autoscaler (VPA) is enabled";

        Value          = { $clusterInfo.properties.workloadAutoScalerProfile.verticalPodAutoscaler.enabled };

        Expected       = $true;

        FailMessage    = "Vertical Pod Autoscaler is disabled, resulting in static resource requests that are often oversized (wasting resources and money) or undersized (causing performance issues and throttling). VPA provides data-driven recommendations to optimize pod resource allocation based on actual usage patterns.";

        Severity       = "Medium";

        Recommendation = "Enable VPA using 'az aks update --resource-group <rg> --name <cluster> --enable-vpa'. Deploy VPA objects with 'updateMode: Auto' or 'Off' for recommendations only. Monitor VPA recommendations and adjust application resource requests/limits accordingly for better resource efficiency.";

        URL            = "https://learn.microsoft.com/azure/aks/vertical-pod-autoscaler";

    },

    @{

        ID             = "AKSRES004";

        Category       = "Resource Management";

        Name           = "KEDA (Event-Driven Autoscaling) Enabled";

        Value          = { $clusterInfo.properties.workloadAutoScalerProfile.keda.enabled };

        Expected       = $true;

        FailMessage    = "KEDA add-on is disabled, preventing event-driven autoscaling for workloads that process queues, messages, or respond to external metrics. This forces reliance solely on CPU/memory-based HPA scaling, which is suboptimal for event-driven architectures and can lead to over-provisioning or delayed scaling during traffic spikes.";

        Severity       = "Low";

        Recommendation = "Enable KEDA using 'az aks update --resource-group <rg> --name <cluster> --enable-keda'. Deploy ScaledObject resources to define event sources (Azure Queue, Service Bus, Kafka, HTTP, etc.) and scaling behavior. KEDA complements HPA by enabling scale-to-zero and event-driven scaling patterns.";

        URL            = "https://learn.microsoft.com/azure/aks/keda-about";

    },

    @{

        ID             = "AKSRES005";

        Category       = "Resource Management";

        Name           = "Node Auto-provisioning or Cluster Autoscaler Configured";

        Value          = { 

            ($clusterInfo.properties.nodeProvisioningProfile.mode -eq "Auto") -or 

            ($clusterInfo.properties.autoScalerProfile -ne $null)

        };

        Expected       = $true;

        FailMessage    = "Neither Node Auto-provisioning (NAP) nor Cluster Autoscaler is enabled, requiring manual node scaling that leads to inefficient resource allocation. NAP uses Karpenter to automatically select optimal VM SKUs based on workload requirements, providing better cost optimization and scaling flexibility than traditional cluster autoscaler.";

        Severity       = "High";

        Recommendation = "Enable Node Auto-provisioning using 'az aks update --resource-group <rg> --name <cluster> --node-provisioning-mode Auto' for Karpenter-based dynamic provisioning. Alternatively, enable Cluster Autoscaler with 'az aks update --enable-cluster-autoscaler'. NAP is recommended for modern workloads with diverse resource requirements.";

        URL            = "https://learn.microsoft.com/azure/aks/node-auto-provisioning";

    }    

)