Cloudwash: An Open Source Tool to Stop Cloud Waste Before It Burns Your Budget

• The problem is universal: Every team running CI/CD pipelines, test automation, or dev/staging environments on popular clouds like AWS, Azure, GCP accumulates forgotten VMs, orphaned disks, dangling NICs, and stale IPs. These ghost resources silently inflate cloud bills — often by 20-30% — and create security surface area that nobody audits.

• Cloudwash is a Python CLI and library that scans your cloud accounts, identifies resources matching configurable cleanup criteria and removes them — across AWS EC2, Azure, GCP, VMware, and Podman containers.

• Designed for safety first: Cloudwash ships with a dry-run mode that generates an HTML report of what would be cleaned, before touching anything. SLA-based time windows ensure recently-created resources are never accidentally deleted. Exception lists let you whitelist critical resources.

• Born at Red Hat to solve real pain: Our CI pipelines provisioned hundreds of cloud VMs daily for testing real products, and the cleanup was manual, error-prone, and costly. Cloudwash automated this entirely, running as a scheduled container on OpenShift, saving thousands of dollars monthly.

• In this talk, I will cover:

  • The architecture: how Cloudwash abstracts multiple cloud providers behind a unified resource model.

  • Live demo: configuring Cloudwash, running a dry-run scan, reviewing the report, and executing cleanup across providers.

  • How we run Cloudwash in production — as a CronJob on OpenShift with DynaConf-based secrets management.

  • Lessons learned building a multi-cloud FOSS tool with a small team, including getting contributions from the community

  • Future roadmap: OpenStack support, cost-estimation in reports, and Kubernetes-native resource cleanup.

  • AI/ML Powered Prediction: How AI/ML would power and predict the cost spent on unnecessary resources etc.

Attendees will leave with a working understanding of how to deploy Cloudwash in their own infrastructure to save unnecessary spends and contribute to the project.