Managed vs unmanaged VPS: the difference, and who does what

Three levels of service

On an unmanaged VPS, the host provides a server that boots, connected to the network, on a machine it maintains. Everything else is yours: the operating system and its updates, software and its configuration, security, monitoring. It is the most common model, designed for teams that run their own servers.

With an assisted VPS, often sold as semi-managed, you keep control of the server without facing a failure or a technical decision alone: you describe what you see, experts diagnose it and tell you what to do, and you apply the fix at your own pace. Depending on the host, that help covers the operating system only or reaches into the application: the depth of the diagnosis is what makes the offer worth having.

On a managed VPS, the host runs the server for you: it applies updates, monitors the server, steps in when an alert fires and restores backups. Proactive management goes further: it also monitors your application and recommends what will make it faster, before you even notice. You focus on your business.

A simple way to tell them apart: assistance is the roadside mechanic you call; full management is the garage that keeps your car serviced.

Who does what, line by line

Task	Unmanaged	Assisted	Managed
Hardware, network, power	Host	Host	Host
Operating system install	Host, from an image	Host, from an image	Host, configured for your use
Security updates	You	You, guided when needed	Host
Installing and tuning services (web, database, cache)	You	You, with recommendations	Host, with you
Monitoring	You	Depends on the offer	Host, with intervention
Diagnosing a failure	You	The host’s experts; you apply the fix	Host
Backups	Depends on the offer	Depends on the offer	Host, restores included
Firewall and access	You	You, guided when needed	Host
Migrating your existing setup	You	You, guided	Often included

The line that matters most is security updates: it is the one that slips when nobody owns it, and unpatched software is one of the first ways into a server.

What to check in a managed hosting contract

Two "managed" offers can have very little in common. These questions clear things up:

1. Scope: the operating system only, or services too (web server, database, email)? Are major versions upgraded, or only patches applied?
2. Monitoring: what is watched (availability, disk, memory, certificates, services), how often, and what happens at night and on weekends?
3. Intervention: does the host act on its own when an alert fires, or wait for your call?
4. Response times: the time to respond (before someone takes the issue in hand) and the time to restore (before the service is back), written into the contract.
5. Backups: frequency, retention, storage location and, above all: are restores included, and how fast are they?
6. What costs extra: interventions beyond a bundle, migrations, configuration changes.
7. Your access: do you keep access to the server, and what kind?
8. Exit terms: at the end of the contract, how do you get your data, your backups and the server documentation back?

Choosing your level

Two questions usually settle it:

• Who in your team knows Linux? A developer at ease on the command line gets the most from a server they control, especially with experts on hand for questions outside their day-to-day. A business with no in-house systems skills gains by handing operations over.
• What does an hour of downtime cost? For a brochure site, an hour offline can be made up. For a shop, a business application or a service your customers rely on, it is counted in orders and in trust: that is where round-the-clock monitoring with intervention earns its place.

Your situation	The level that fits
You build, you deploy, and you want to stay in control	Assisted
Your team already runs servers, with an on-call rotation	Unmanaged, or assisted for a second pair of eyes
A critical application, with no in-house systems skills	Managed
Uptime commitments written into your own contracts	Managed

Cost compares with what it replaces: the time someone in your team spends on the server, and the price of an outage. Managed hosting costs more than a bare VPS because it brings expert time and commitments; it pays off as soon as that time and those commitments are what you are missing.

Changing level along the way

The level of service is not fixed. Many projects start on a server that one team member runs with assistance, then move to full management once the application becomes critical or the team grows. The handover is prepared: an inventory of what runs on the server, documentation of its settings, and an agreed switchover time, especially when management comes with a migration to a new server.