grapes-ecs 0.1.0

Single pane TUI for monitoring AWS ECS cluster health

Grapes

Warning This project is an experiment in AI-assisted development. The code was generated without human review and should not be trusted for production use. That said, Grapes is read-only and makes no modifications to your AWS resources, so there's no harm in trying it!

A single-pane TUI (Text User Interface) for monitoring AWS ECS cluster health. Built with Python and Textual.

Features

• Cluster Selection: View all accessible ECS clusters and select one to monitor
• Real-time Monitoring: View cluster, service, task, and container status at a glance
• Health Indicators: Visual health status (✓ Healthy, ⚠ Warning, ✗ Unhealthy, ? Unknown)
• Container Metrics: CPU and memory utilization from CloudWatch Container Insights
• Deployment Tracking: See all deployments (PRIMARY, ACTIVE, etc.) with task definition versions
• AWS Console Integration: One-key access to open resources in AWS Console
• Keyboard Navigation: Fast, keyboard-driven interface
• Dark Mode: Easy on the eyes for extended monitoring sessions

Requirements

• Python 3.11+
• AWS credentials configured (via environment, ~/.aws/credentials, or IAM role)
• CloudWatch Container Insights enabled (optional, for CPU/memory metrics)

Installation

Install from PyPI:

pip install grapes-ecs

Or using uv:

uv tool install grapes-ecs

Then run:

grapes

Development Installation

# Clone the repository
git clone <repository-url>
cd grapes

# Install dependencies
uv sync

# Run the application
uv run grapes

Configuration

Create a config.toml file in your current directory or at ~/.config/grapes/config.toml:

[cluster]
name = "my-ecs-cluster"      # Optional: ECS cluster name (if omitted, select from list)
region = "us-east-1"         # Required: AWS region
profile = "default"          # Optional: AWS profile name

[refresh]
interval = 30                # Optional: API poll interval in seconds (default: 30)
task_definition_interval = 300  # Optional: Task def cache TTL in seconds (default: 300)

If you omit the name field, Grapes will show a list of all clusters in the region, and you can select one to view.

Run with a specific config file:

grapes -c /path/to/config.toml

Keyboard Shortcuts

Key	Action
↓/↑	Navigate list
Enter	Select cluster / View service details
Esc	Go back (cluster list / service list)
O	Open in AWS Console
R	Force refresh data
D	Toggle debug console
Q	Quit

Required IAM Permissions

The following IAM permissions are required:

{
  "Version": "2012-10-17",
  "Statement": [
    {
      "Effect": "Allow",
      "Action": [
        "ecs:ListClusters",
        "ecs:DescribeClusters",
        "ecs:ListServices",
        "ecs:DescribeServices",
        "ecs:ListTasks",
        "ecs:DescribeTasks",
        "ecs:DescribeTaskDefinition"
      ],
      "Resource": "*"
    },
    {
      "Effect": "Allow",
      "Action": [
        "cloudwatch:GetMetricData",
        "cloudwatch:GetMetricStatistics"
      ],
      "Resource": "*"
    }
  ]
}

Enabling Container Insights

For CPU and memory metrics, enable Container Insights on your cluster:

aws ecs update-cluster-settings \
  --cluster my-cluster \
  --settings name=containerInsights,value=enabled \
  --region us-east-1

If Container Insights is not enabled, Grapes will display a warning and show - for metrics.

Health Status

Grapes determines health based on ECS container health checks:

• ✓ Healthy: All containers report healthy
• ⚠ Warning: Some containers unhealthy or desired != running count
• ✗ Unhealthy: All containers unhealthy or no running tasks
• ? Unknown: Health checks not configured

No fallback logic is used - only actual container health check results are displayed.

Metrics Display

When Container Insights is enabled:

• CPU: Displayed as usage% / limit_units (e.g., 12% / 512)
• Memory: Displayed as usedM / limitM (e.g., 256M / 1024M)

Metrics are fetched from CloudWatch and may have 1-2 minutes of lag. Missing or stale metrics are always shown as -.

Development

# Install dev dependencies
uv sync

# Run linting
uv run ruff check .

# Run tests
uv run pytest

License

MIT