Wasteless

Open-source cloud cost optimization. Detect idle EC2 instances. Remediate with one click.

---

Quick Start

Requirements: Docker, Python 3.11+, AWS credentials (aws configure)

git clone https://github.com/wastelessio/wasteless.git
cd wasteless
./install.sh          # Installs everything (backend + UI + DB)

source ~/.zshrc
wasteless             # Start the web UI

Open http://localhost:8888

Then collect data and detect waste:

source venv/bin/activate
python3 src/collectors/aws_cloudwatch.py   # Collect metrics
python3 src/detectors/ec2_idle.py          # Detect waste

---

Architecture

┌─────────────────────────────────────────────────────────┐
│                    AWS Account                           │
│         CloudWatch API  ·  Cost Explorer  ·  EC2        │
└──────────────────────┬──────────────────────────────────┘
                       │ boto3
┌──────────────────────▼──────────────────────────────────┐
│                     wasteless                            │
│                                                          │
│  collectors/    →   Fetch CloudWatch metrics             │
│  detectors/     →   Identify idle EC2 instances          │
│  remediators/   →   Stop / terminate instances           │
│  trackers/      →   Verify actual savings                │
└──────────────────────┬──────────────────────────────────┘
                       │ psycopg2
┌──────────────────────▼──────────────────────────────────┐
│                     PostgreSQL                           │
│    ec2_metrics · waste_detected · recommendations        │
│    actions_log · rollback_snapshots · savings_realized   │
└──────────────────────┬──────────────────────────────────┘
                       │
┌──────────────────────▼──────────────────────────────────┐
│                      ui/  (FastAPI)                      │
│         Dashboard · Recommendations · History            │
│                    :8888                                 │
└─────────────────────────────────────────────────────────┘

---

Repository Structure

wasteless/
├── src/
│   ├── collectors/         # CloudWatch metrics collection
│   ├── detectors/          # EC2 idle waste detection
│   ├── remediators/        # Stop / terminate execution
│   ├── trackers/           # Savings verification
│   └── core/               # Database, safeguards
├── ui/                     # FastAPI web dashboard
│   ├── main.py
│   ├── templates/
│   ├── utils/
│   ├── install.sh
│   └── start.sh
├── sql/                    # Database schema + migrations
├── config/
│   └── remediation.yaml   # Safeguards and policies
├── docker-compose.yml      # PostgreSQL (+ Metabase optionnel)
└── requirements.txt

---

Configuration

Environment variables

Copy .env.template to .env:

Variable	Default	Description
AWS_REGION	us-east-1	AWS region
AWS_ACCESS_KEY_ID	(required)	AWS access key
AWS_SECRET_ACCESS_KEY	(required)	AWS secret key
DB_HOST	localhost	PostgreSQL host
DB_PORT	5432	PostgreSQL port
DB_NAME	wasteless	Database name
DB_USER	wasteless	Database user
DB_PASSWORD	(required)	Database password

Remediation policy

Edit config/remediation.yaml:

Setting	Default	Description
auto_remediation.enabled	false	Enable autonomous execution
auto_remediation.dry_run_days	7	Mandatory dry-run period
protection.min_instance_age_days	30	Ignore instances younger than N days
protection.min_idle_days	14	Must be idle for N+ days
protection.min_confidence_score	0.80	Minimum detection confidence
protection.max_instances_per_run	3	Max instances stopped per run

Whitelist instances by ID or tag:

whitelist:
  instance_ids:
    - i-0123456789abcdef0
  tags:
    Environment: production

AWS IAM permissions

Minimum required policy:

{
  "Action": [
    "cloudwatch:GetMetricStatistics",
    "cloudwatch:ListMetrics",
    "ec2:Describe*",
    "ec2:StopInstances",
    "ce:GetCostAndUsage"
  ]
}

---

Safeguards

Before executing any action, Wasteless validates 7 conditions:

1. Auto-remediation enabled in config
2. Instance not in whitelist
3. Instance age >= 30 days
4. Detection confidence >= 80%
5. Idle duration >= 14 consecutive days
6. Current time in allowed schedule window
7. Instances stopped this run < max limit

If any check fails → action aborted and logged.

---

Features

Feature	Description
Detection	CPU / network analysis with confidence scoring
Recommendations	Stop / terminate / downsize actions
Dry-Run Mode	Test safely before any AWS action
Auto-Sync	Background sync every 5 min with AWS
Action History	Full audit trail with rollback snapshots
Cloud Inventory	Live EC2 inventory across regions
Whitelist	Exclude instances from recommendations
Savings Tracking	Verified actual savings via Cost Explorer

---

API (UI)

Endpoint	Method	Description
/	GET	Home dashboard
/recommendations	GET	Pending recommendations
/history	GET	Action history
/cloud-resources	GET	EC2 inventory
/settings	GET	Configuration
/api/metrics	GET	JSON metrics
/api/actions	POST	Approve / reject
/api/config	POST	Update config
/api/whitelist	POST	Add to whitelist
/api/sync-aws	POST	Trigger manual sync

---

Compatibility

OS	Status
macOS	Supported
Linux	Supported
Windows (WSL2)	Supported
Windows (native)	Not supported

---

Development

# Backend tests
python tests/test_end_to_end.py

# UI hot reload
cd ui && ./start.sh

# UI tests
cd ui && python run_tests.py

---

Roadmap

• EC2 idle detection and remediation
• Web dashboard (FastAPI)
• Dry-run mode and safeguards
• Savings verification
• RDS / EBS / S3 detection
• Multi-account AWS support
• Slack / Teams notifications
• Azure and GCP support

---

License

Apache 2.0

---

Links

• Issues: GitHub Issues
• Contact: wasteless.io.entreprise@gmail.com