Your Outscale dev environment is bigger than you think

You think your dev environment costs €1,500 a month. You're probably wrong by half.

Dev environments are where Outscale bills hide. Production has metrics, dashboards, and on-call rotations. Production gets watched. Dev gets used 8 hours a day by humans and ignored the other 16. It also gets used by every "let me try something" experiment that didn't get cleaned up, every staging cluster that should have been a single instance, every snapshot from a project that ended six months ago.

Why dev grows silently

The structural reasons are predictable:

• Nobody owns dev. Production has SREs. Dev has whoever set it up, who left, and now nobody.
• Cost ≠ pain. Production downtime is felt instantly. Dev cost is felt at month-end, by someone in finance, who emails someone in engineering.
• Cleanup has no incentive. Provisioning is a 5-minute reward (the thing works). Decommissioning is a 30-minute chore with no observable upside.
• Defaults are aggressive. The default volume size is 100 GiB. The default schedule is 24/7. The default snapshot retention is forever. None of these match what dev actually needs.

The four cost drivers

For most dev environments, 90% of the cost lives in four buckets:

1. Running compute when nobody's working. A dev VM running 24/7 on a typical Mon-Fri 8-18 schedule wastes 70% of its compute hours. Across 20 VMs that's not "let's optimize someday" — it's the single biggest line item you'll see when you tag dev resources and look.

2. Compute over-sized for the workload. The VM was sized for the day someone ran a load test. It's been at 5% CPU since. Scheduling cuts the time, right-sizing cuts the rate. Both compound.

3. Orphan volumes. Volumes that were attached to instances that got terminated but the volume didn't. They're still billing at full GiB-month rate, doing nothing. Easy to enumerate via the API; rarely enumerated in practice.

4. Snapshot accumulation. Every snapshot is small. A few hundred snapshots is not. Snapshot retention is the canonical "set it and forget it" disaster — works fine until you check the bill and find six months of weekly backups for an environment that's been gone for three.

How to actually find them

Three tools, in order:

Tags. If your dev resources are tagged (environment=dev, team=...), you can filter Cost Explorer by tag and see the dev sub-bill. If they're not tagged, that's job zero — you can't manage what you can't measure. Apply the tags via the API in 30 minutes; the cost analysis becomes possible immediately.

Cost Explorer with daily granularity. Look at dev cost over the last 30 days, by day. Do you see weekend dips? You should. If the line is flat across Saturdays and Sundays, your environment runs 24/7. There's your scheduling target.

Resource inventory cross-checked against running state. Pull every BSU volume on the account, filter to those with no attached instance. Pull every snapshot, filter to those older than 90 days. These are both single API calls. The list is your cleanup queue.

The 50-hour rule

A normal work week is 50 hours of activity (8am-6pm × 5 days). A normal dev environment runs 168 hours. That's a 70% gap, every week, every month, forever — until you put a schedule on it.

The math is simple. The objection isn't math, it's "what if someone needs it on a Saturday." Real answer: they almost never do, and when they do, they can start the environment via a button in 30 seconds. The cost of "always available" is paid every week. The cost of "manually start when needed" is paid maybe once a quarter.

A weekly habit

The easiest way to keep dev costs under control is to look at them weekly. 5 minutes, every Monday. What's running? What's growing? What's still here from last quarter?

Most teams do this monthly because that's when the invoice arrives. By then it's already been a month of waste. Weekly catches problems while they're a quarter the size.

If you have Vextnd connected to your account, dev costs sit in Cost Explorer with the relevant filters saved. Five minutes a week — and the savings compound the same way the waste did before you started looking.