A Smarter Way to Manage EC2 for Business-Hours Apps
Resize, reboot, and stay online. This simple change cut EC2 costs by over $43K a year—with no impact to teams or tools.
Cloud cost optimization usually starts with turning things off. But for many teams, especially those managing integrated workloads and legacy apps, shutting down EC2 instances breaks more than it saves. In this post, I’ll walk through a smarter way to reduce EC2 spend for business-hours systems—by resizing them instead. One reboot. No false alarms. No team disruption. Over $43,000 saved in one year.
1. Why This Matters
Many environments—especially dev, test, and staging—run 24/7. But they’re only used during business hours. Shutting them down entirely overnight or on weekends sounds great until endpoint monitoring, backup agents, or security tools start throwing alerts. Resizing instead of shutting down avoids those problems while still delivering savings.
2. The Approach
A scheduled AWS Lambda function, triggered by EventBridge, finds EC2 instances tagged with DynamicInstanceScaling = true. At 7 PM Pacific, it stops each instance, resizes it to a smaller type like t3.medium, and starts it again. At 4 AM Pacific, the function resizes them back to their original types using the PreferredInstanceType tag. The reboot takes only a few minutes and happens when nobody’s using the systems.
3. Why It Works
This method keeps systems running while lowering compute costs. Monitoring tools stay connected. Agent-based security scans still work. Backups don’t fail. Teams don’t notice anything different. It also complements existing Compute Savings Plans. When workloads shrink overnight, unused capacity can shift to other systems—extending the value of committed spend.
4. Real Impact
In one deployment, over 100 EC2 instances were enrolled in this schedule. Most were resized down to t3.medium after hours and back up to their normal types each weekday morning. This resulted in monthly savings of more than $3,600, which added up to over $43,000 in one year. Those savings came on top of a 3-year Compute Savings Plan. Teams didn’t need to update a single workflow. Systems stayed up, integrations held, and tools continued reporting normally.
This approach isn’t limited to non-production. It works for any business-hours workload—including production apps used by internal teams. If a system is quiet overnight, it doesn’t need full compute power. A quick reboot and resize is enough.
5. The Bottom Line
Business-hours systems don’t need full capacity 24/7. Resizing instead of shutting down lets you reduce costs without adding operational noise. It works with your existing tools, agents, and cloud commitments. The impact is minimal. The savings are real.
How are you handling EC2 scaling in your business-hours environments? I’d love to compare notes and see what’s working for your team.