Loading
2 October 2026 · 4 min read
When stopped servers are not billed for CPU and RAM, environments that sleep outside business hours win back a large share of your spend. Here is a worked example.
For many SaaS companies, the cloud bill is one of the fastest-growing expenses. The most common reason is that environments used only during business hours run for the entire month. Hourly billing targets that waste directly.
In the traditional approach, capacity is planned up front and stays fixed. Growing demand requires manual intervention, and the result is either over-provisioning or a system that struggles at peak. Development, test and demo environments are often left running 24/7 because stopping and starting them feels like a hassle.
On Bulutlan Edge Cloud, the resources you use are billed by the hour, with no hidden costs. The basic rules are:
Prices are in USD and exclude VAT.
In an example from our infrastructure partner RNG Technology, the CPU and RAM cost of a server running 730 hours is 50 USD. When the same server runs 12 hours a day and stays off on weekends, monthly usage drops to about 240 hours and the cost to 16.43 USD. That is a difference of about 70%.
Let's run the same numbers with the Z2 profile (2 CPU, 4 GB RAM) from the Bulutlan price list:
The math is simple: 24.82 x 240 / 730 ≈ 8.16 USD. For a six-server test environment, the monthly difference is about 100 USD. Disks and IPs are not included in these figures and continue to be billed while the server is stopped.
You can try your own configuration in the Pricing Calculator. Its monthly running hours setting applies only to server profiles billed by the hour.
Not every workload has to run 24/7. In SaaS companies, the best candidates for stopping are:
Even if production keeps running 24/7, moving just these environments to a business-hours schedule makes a visible difference on the bill.
Hourly billing helps on the production side too. With Auto-Scale, you define minimum, desired and maximum capacity. When load spikes, extra servers come online automatically; when it drops, they shut down. Instead of a fixed fleet sized for your busiest moment, you pay for extra resources only when you need them.
With infrastructure as code, you can go one step further. A test environment defined in Terraform can be removed entirely when the work is done and rebuilt in minutes when needed. In that case you stop paying for disks and IPs as well as servers. You only need a separate storage layer for data that has to persist.
Stopping a server does not delete data; disks and the data on them are kept. Still, check a few things before you move to a stop schedule:
Hourly billing ties your cloud cost to how you actually work. Stopping environments that do not need to run outside business hours and using Auto-Scale for peaks are the most direct ways to lower cost without giving up performance. Full prices are on the Pricing page.