Istio Usage Collector

This tool collects information from your Kubernetes clusters to analyze resource usage and help estimate the cost and resources required for migrating to ambient mesh.

Description

The Istio Usage Collector is a Go implementation of the gather-cluster-info.sh script. It gathers non-sensitive information about your Kubernetes cluster, including:

• Node information (instance type, region, zone, CPU, memory)
• Namespace information
• Pod and container counts
• Resource requests and usage
• Istio sidecar information

This data is collected into a JSON or YAML file that can be used for further analysis through our detailed migration estimator tool.

Installation

Downloading release

You can download the latest release for your system by passing

curl -sL https://storage.googleapis.com/istio-usage-collector/install.sh | sh -

You can also specify a release version, os, and arch to download specific version(s) by passing the relevant flags

curl -sL https://storage.googleapis.com/istio-usage-collector/install.sh | sh -s -- --os darwin --arch arm64 --version latest

Building from source

1. Clone the repository:

   git clone https://github.com/solo-io/istio-usage-collector.git
   cd istio-usage-collector
   

2. Build the binary:

   make build
   

Usage

./istio-usage-collector [subcommand] [flags]

Flags

• --version or -v: Print the version information.
• --hide-names or -n: Hide the names of the cluster and namespaces using a hash.
• --continue or -c: If the script was interrupted, continue processing from the last saved state.
• --context or -k: Kubernetes context to use (if not set, uses current context).
• --output-dir or -d: Directory to store output files (default: current directory).
• --format or -f: Output format (json, yaml/yml) (default: json).
• --output-prefix or -p: Custom prefix for output files (default: cluster name).
• --help or -h: Show help message.
• --no-progress: Disable the progress bar.
• --debug: Enable debug logs.

Example

# Use the current context with default JSON output
./istio-usage-collector

# Get the version information
./istio-usage-collector version

# Use a specific context - this would scan `my-cluster` and be saved as ./my-cluster.json
./istio-usage-collector --context my-cluster

# Output in YAML format - this would be saved as ./<cluster>.yaml
./istio-usage-collector --format yaml

# Specify a custom output file prefix (name) - this would be saved as ./prod-cluster.json
./istio-usage-collector --output-prefix prod-cluster --format json

# Hide sensitive names and save to custom location - this would be saved as /reports/<hashed-cluster>.yaml
./istio-usage-collector --hide-names --output-dir /reports --format yaml

# Continue an interrupted collection
# Note that in order to successfully continue, the original flags must be passed as well.
./istio-usage-collector --continue

Output

The tool generates output files containing the collected information:

• JSON format: <cluster-name>.json (default)
• YAML format: <cluster-name>.yaml

JSON Output Structure

When using the default JSON format, the file has the following structure:

{
  "name": "cluster-name",
  "namespaces": {
    "namespace1": {
      "pods": 10,
      "is_istio_injected": true,
      "resources": {
        "regular": {
          "containers": 15,
          "request": {
            "cpu": 2.5,
            "memory_gb": 4.0
          },
          "actual": {
            "cpu": 1.2,
            "memory_gb": 2.1
          }
        },
        "istio": {
          "containers": 10,
          "request": {
            "cpu": 1.0,
            "memory_gb": 1.5
          },
          "actual": {
            "cpu": 0.5,
            "memory_gb": 0.8
          }
        }
      }
    }
  },
  "nodes": {
    "node1": {
      "instance_type": "m5.large",
      "region": "us-east-1",
      "zone": "us-east-1a",
      "resources": {
        "capacity": {
          "cpu": 2,
          "memory_gb": 8
        },
        "actual": {
          "cpu": 1.5,
          "memory_gb": 6.0
        }
      }
    }
  },
  "has_metrics": true
}