On this page
Buying a server is the start of hosting. Managed hosting is the work that happens after the login is handed over: patches, backups, capacity, and a person who answers when the site is down at 02:00.
Teams often discover the gap during the first incident. The host replaced the disk. Nobody restored the application. Both parties did their job. Neither job was the one the business needed.
The four things that should be in the agreement
- Monitoring that checks the application, not only ping. A server that answers ICMP while the checkout page returns 500 is not up.
- Patching with a window and a rollback. Unattended updates on a database host are not a patch policy.
- Backups with a tested restore. A backup that has never been restored is a file on a disk.
- Response with a named severity, a clock, and a channel. "Email us" is not a 24/7 path.
Ask who is allowed to reboot, who holds the DNS, and who talks to the application vendor. Those three answers tell you whether you have managed hosting or a rented machine with a phone number.
What stays with your team
Managed hosting does not replace product decisions. You still choose what ships, which data you collect, and which compliance programme you are in. The operator should make those choices durable: the release you approved is the release that is running, and the change is written down.
If your current host sends an invoice and a status page, and nothing else, the next incident will explain the difference. We can walk through your stack and say plainly what is covered today.
