Skip to content

CRD Commands

Comprehensive guide for working with Custom Resource Definitions (CRDs) using K8s Inventory CLI.

Overview

The CRD commands provide powerful capabilities to discover, analyze, and inventory Custom Resource Definitions in your Kubernetes cluster.

Available Commands

List CRDs

Display all CRDs in your cluster with comprehensive metadata:

k8s-datamodel crd list

Filtering Options

Filter CRDs by API group: 

k8s-datamodel crd list --group networking.k8s.io
k8s-datamodel crd list --group operators.coreos.com

Filter by resource kind: 

k8s-datamodel crd list --kind VirtualService
k8s-datamodel crd list --kind Subscription

Filter by scope (Namespaced/Cluster): 

k8s-datamodel crd list --scope Namespaced
k8s-datamodel crd list --scope Cluster

Get CRD Details

Retrieve detailed information about a specific CRD:

k8s-datamodel crd get certificaterequests.cert-manager.io

This command provides: - Complete CRD specification - All available versions - Conversion strategy - Storage version - Categories and short names - Age and creation timestamp

Count CRD Instances

Count how many custom resource instances exist for each CRD:

k8s-datamodel crd count

With filtering: 

k8s-datamodel crd count --group cert-manager.io
k8s-datamodel crd count --scope Cluster

CRD Analysis Process

flowchart LR
    subgraph "Discovery Phase"
        API["⚡ K8s API"]
        FETCH["📥 Fetch CRDs"]
        FILTER["🔍 Apply Filters"]
    end

    subgraph "Analysis Phase"
        PARSE["🔍 Parse Spec"]
        FRAMEWORK["🏷️ Detect Framework"]
        CATEGORIZE["📊 Categorize"]
        COUNT["🔢 Count Instances"]
    end

    subgraph "Output Phase"
        FORMAT["🎨 Format Results"]
        DISPLAY["💻 Display"]
        EXPORT["📁 Export"]
    end

    API --> FETCH
    FETCH --> FILTER
    FILTER --> PARSE
    PARSE --> FRAMEWORK
    FRAMEWORK --> CATEGORIZE
    CATEGORIZE --> COUNT
    COUNT --> FORMAT
    FORMAT --> DISPLAY
    FORMAT --> EXPORT

    style API fill:#fff3e0
    style DISPLAY fill:#e8f5e8
    style EXPORT fill:#e8f5e8

Understanding CRD Information

Framework Detection

K8s Inventory CLI automatically detects how CRDs were deployed:

  • Helm: CRDs installed via Helm charts
  • OLM: CRDs managed by Operator Lifecycle Manager
  • Manual: CRDs applied directly via kubectl/API

CRD Categories

CRDs are automatically categorized based on their metadata: - API extensions - Networking resources - Security policies - Storage classes - Monitoring resources

Instance Analysis

The tool can analyze actual custom resource instances: - Count instances per namespace - Identify unused CRDs - Track resource creation patterns - Monitor resource age distribution

CRD Relationship Mapping

erDiagram
    CRD ||--o{ CUSTOM_RESOURCE : "defines"
    CRD ||--|| OPERATOR : "managed_by"
    CRD {
        string name
        string group
        string kind
        string scope
        string framework
        int instance_count
    }

    CUSTOM_RESOURCE {
        string name
        string namespace
        string status
        datetime created
    }

    OPERATOR {
        string name
        string namespace
        string framework
        string health_status
        int replicas
    }

    NAMESPACE ||--o{ CUSTOM_RESOURCE : "contains"
    NAMESPACE {
        string name
        string status
    }

    FRAMEWORK ||--o{ CRD : "deploys"
    FRAMEWORK {
        string name
        string description
    }

Advanced Filtering

Multiple Filters

Combine multiple filters for precise results: 

k8s-datamodel crd list --group networking.k8s.io --scope Namespaced

Pattern Matching

Use wildcards in group names: 

k8s-datamodel crd list --group "*.coreos.com"

Output Formats

All CRD commands support multiple output formats:

Table Format (Default)

k8s-datamodel crd list --output table
Human-readable grid with key information.

Rich Format

k8s-datamodel crd list --output rich
Enhanced terminal output with colors and styling.

JSON Format

k8s-datamodel crd list --output json
Machine-readable format perfect for scripting.

YAML Format

k8s-datamodel crd list --output yaml
Structured format for configuration management.

Use Cases

Cluster Migration Planning

# Identify all CRDs that need to be migrated
k8s-datamodel crd list --output yaml > crds-inventory.yaml

# Check which CRDs have active instances
k8s-datamodel crd count --output json > crd-usage.json

Security Auditing

# List all cluster-scoped CRDs (potential security impact)
k8s-datamodel crd list --scope Cluster

# Find CRDs from specific vendors
k8s-datamodel crd list --group "*.example.com"

Operator Management

# Find all OLM-managed CRDs
k8s-datamodel crd list | grep "OLM"

# Analyze CRDs by framework
k8s-datamodel crd list --output json | jq '.[] | select(.framework == "Helm")'

Cleanup Operations

# Find CRDs with zero instances
k8s-datamodel crd count | grep "0 instances"

# Identify old/deprecated CRDs
k8s-datamodel crd list | sort -k6  # Sort by age

Troubleshooting

Permission Issues

If you encounter permission errors: 

# Test cluster connection
k8s-datamodel cluster test-connection

# Check your RBAC permissions
kubectl auth can-i get customresourcedefinitions

Large Clusters

For clusters with many CRDs, use filtering to improve performance: 

# Instead of listing all CRDs
k8s-datamodel crd list --group specific.domain.com

Output Formatting

Use --verbose flag for debugging: 

k8s-datamodel crd list --verbose

Integration Examples

CI/CD Pipeline Integration

#!/bin/bash
# Export CRD inventory for compliance reporting
k8s-datamodel crd list --output json > artifacts/crd-inventory.json

# Check for unauthorized CRDs
UNAUTHORIZED=$(k8s-datamodel crd list --group "untrusted.com" --output json)
if [[ "$UNAUTHORIZED" != "[]" ]]; then
  echo "Unauthorized CRDs detected!"
  exit 1
fi

Monitoring Integration

# Generate metrics for monitoring systems
k8s-datamodel crd count --output json | \
  jq -r '.[] | "\(.name) \(.instance_count)"' | \
  while read name count; do
    echo "crd_instances{name=\"$name\"} $count"
  done