Odigos Uninstrumented Sources Tool

A command-line tool that identifies and instruments Kubernetes workloads that are not currently instrumented by Odigos. The tool connects to your Odigos installation, queries for workloads missing instrumentation, and can automatically enable instrumentation for them.

What It Does

1. Checks if kubectl is available in your PATH
2. Establishes a port-forward to the Odigos UI service (svc/ui in the specified namespace)
3. Connects to the Odigos GraphQL API at http://localhost:3000/graphql
4. Fetches all namespaces from the cluster
5. Filters out ignored namespaces (default: kube-system, kube-public, default)
6. Fetches workloads for each remaining namespace
7. Identifies workloads that are not instrumented
8. Optionally exports uninstrumented workloads to a CSV file
9. Updates the instrumentation status for uninstrumented workloads (unless --dry-run is specified)

Requirements

• Python 3.10+
• kubectl - Must be installed and available in PATH
• Kubernetes cluster access - kubectl must have access to the Odigos namespace (specified via --odigos-namespace)
• Odigos installation - The Odigos UI service must be running in the cluster

Installation

# Clone the repository
git clone <repository-url>
cd add-uninstrumented-sources

# Using uv (recommended)
uv sync

# Or using pip
pip install -e .

Usage

python odigos_instrument.py [OPTIONS]

Options

Option	Description
--odigos-namespace NAMESPACE	Required. The namespace where Odigos is installed (e.g., odigos-system)
--log-level	Set logging level: DEBUG, INFO, WARNING, ERROR, CRITICAL (default: INFO)
--log-file FILE	Write logs to a file in addition to console
--export-csv FILE	Export uninstrumented workloads to a CSV file
--ignore-namespaces NAMESPACE [NAMESPACE ...]	List of namespaces to ignore when fetching workloads (default: kube-system kube-public default)
--dry-run	Preview changes without updating instrumentation
--help	Display help message

Examples

Basic Usage

Run with default INFO logging level (namespace is required):

python odigos_instrument.py --odigos-namespace odigos-system

Dry Run (Preview Only)

See what would be updated without making changes:

python odigos_instrument.py --odigos-namespace odigos-system --dry-run

Export to CSV

Export uninstrumented sources to a CSV file without updating:

python odigos_instrument.py --odigos-namespace odigos-system --export-csv uninstrumented.csv --dry-run

Export to CSV and update instrumentation:

python odigos_instrument.py --odigos-namespace odigos-system --export-csv sources.csv

Ignoring Namespaces

By default, the tool ignores kube-system, kube-public, and default namespaces. You can customize this:

# Ignore only kube-system
python odigos_instrument.py --odigos-namespace odigos-system --ignore-namespaces kube-system

# Ignore multiple custom namespaces
python odigos_instrument.py --odigos-namespace odigos-system --ignore-namespaces kube-system monitoring logging

Logging Options

Run with DEBUG logging for detailed output:

python odigos_instrument.py --odigos-namespace odigos-system --log-level DEBUG

Run with WARNING level to suppress info messages:

python odigos_instrument.py --odigos-namespace odigos-system --log-level WARNING

Log to a file:

python odigos_instrument.py --odigos-namespace odigos-system --log-file /path/to/logfile.log

Combine DEBUG level with file logging:

python odigos_instrument.py --odigos-namespace odigos-system --log-level DEBUG --log-file debug.log

CSV Output Format

When using --export-csv, the tool creates a CSV file with the following columns:

Column	Description
namespace	Kubernetes namespace of the workload
name	Name of the workload
kind	Workload type (e.g., Deployment, StatefulSet, DaemonSet)

License

See LICENSE file for details.