A VPS grows with your project: the plan that was enough at launch becomes tight once traffic, the database or the number of sites increases. Moving to a higher tier is done without reinstalling: your data comes with you, your operating system stays in place and you do not start from scratch.
When a higher tier is justified
free -hshows swap being used permanently.- The kernel kills processes for lack of memory (
dmesg | grep -i oom). - The load average sustainably exceeds the number of available vCPUs.
- The disk stays above 85% full despite cleanup.
- Response time degrades at peak hours.
A single signal is not always enough; several signals together indicate a real need.
What moving up changes
- VPS-1G — Starter: 1 vCPU, 1 GB RAM, 30 GB NVMe SSD.
- VPS-2G — Business: 1 vCPU, 2 GB RAM, 40 GB NVMe SSD.
- VPS-3G — Performance: 2 vCPU, 3 GB RAM, 50 GB NVMe SSD.
- VPS-4G — Advanced: 2 vCPU, 4 GB RAM, 80 GB NVMe SSD.
- VPS-5G — Ultra: 3 vCPU, 5 GB RAM, 100 GB NVMe SSD.
Each tier adds memory and vCPUs, and the disk capacity follows. Memory is what most often changes the picture: it prevents swapping, and swapping is what makes a server slow while its processor is barely used.
Before migrating: back up first
A migration takes preparation. Take a full backup of your files and your database, then verify that it can be read. Write down the installed versions, running services, scheduled tasks and associated domain names. Finally, plan a maintenance window, even a short one: any intervention on a production server deserves to be scheduled.
How to request the migration
- Open a ticket from your client area, selecting the service concerned.
- State the target tier and the server's actual use case.
- The support team confirms the procedure, the intervention window and the agreed date.
- Avoid changing the server configuration during the operation.
Verify after the migration
nproc free -h df -h systemctl --failed
Then check your sites, databases, HTTPS certificates and scheduled tasks. Compare response time before and after: it is the most telling indicator for validating the operation.
What if the new tier is not enough?
Monitor the same indicators for a few days. If swap comes back or the load stays high, the next tier is probably the right answer. If on the contrary everything stays very low, you may consider moving back down: the important thing is to decide from measurements, not from an impression.
