Do We Still Need This? - Automating RDS Cleanup with Data-Driven Triage

RDS workloads are one of the easiest places for cloud waste to hide—and one of the hardest to clean up.

Low vCPU utilization. Sporadic connections. No clear owner. Still running 24/7.

In this hands-on FinOps Fix clinic, we’ll break down how high-performing teams are automating the full lifecycle of RDS cleanup—using real usage signals, not assumptions.

This session will be led by Dieter Matzion, Sr. Cloud Governance Engineer at Roku, sharing real-world strategies for managing database sprawl at scale. Together, we’ll walk through:

• How to analyze 90-day vCPU and connection patterns to identify underutilized and idle databases

• How to differentiate between truly idle workloads and “quiet but critical” systems

• What needs to be in place before automating cleanup

• How to drive developer engagement during triage—without endless Slack threads

• A live walkthrough of how the workflow runs in Wiv.ai, including the internal automation tracker used to manage engagement

This isn’t a dashboard demo.
It’s a technical walkthrough focused on how automation actually gets built and run in production.

If RDS sprawl is quietly eating your budget, this clinic will show you how to fix it—or flag it—with data and automation.