comparison Hosting
Managed vs unmanaged hosting
“Managed” describes which jobs a provider performs, not whether the machine is shared, virtual or dedicated.
What it means
Managed hosting means a provider takes responsibility for specified operational tasks. Unmanaged hosting usually leaves more of those tasks to the customer. The labels do not define a universal checklist. The contract and product documentation determine who patches systems, monitors failures, maintains backups and responds to incidents. Shared responsibility
A VPS may be managed or unmanaged; so can a dedicated machine or a web application platform. Management level is a different axis from server size, physical tenancy and virtualization. Do not compare “managed” with “VPS” as if one excludes the other.
Try it
Simulation — assign work before rehearsing an incident. Select an agreement, inspect a task, and save a proposed task owner with Assign task locally. Rehearse incident then routes the next action using that saved agreement. Merely selecting an owner does not change the agreement.
Responsibility and incident rehearsal
Assign a task, then see who must respond to a synthetic incident.
Selecting an assignee is a draft. Assign task locally changes the agreement; rehearsing an incident uses that saved agreement.
| Item | State / result | Why it matters |
|---|---|---|
| OS updates | provider | Confirm scope, access and verification. |
| Application updates | owner | OS management does not transfer this task. |
| Credentials | owner | Confirm scope, access and verification. |
| Monitoring | provider | Confirm scope, access and verification. |
| Backup creation | provider | Confirm scope, access and verification. |
| Restoration | unknown | A stored backup does not establish recovery. |
Invented agreements, not a review of a real contract or proof an incident is fixed.
About this model
An agreement must name tasks, authority, escalation and verification. Provider responsibility for OS patches does not cover application patches automatically. Backup creation and restoration are distinct tasks. Shared responsibility needs a named handoff; unknown or absent ownership remains a gap.
Read the article’s primary sources for protocol rules and implementation-specific limits. This scene has no network access.
In the managed example, OS updates belong to the provider, but application updates belong to the website owner. Rehearse an application patch incident: provider-managed OS maintenance is not proof that the application has been patched. Next choose a restore incident. Backup creation is assigned to the provider, while restoration is unknown; a backup task cannot silently fill that gap.
| Agreement observation | Result of the rehearsal | Evidence still required outside the scene |
|---|---|---|
| Provider owns OS, owner owns application | Application incident routes to owner | Supported patch procedure and post-patch verification |
| Provider creates backups, restore owner unknown | Ownership gap remains open | Who can authorize, perform and verify a restore |
| Shared task | Named handoff is needed | Actor, access, escalation and verifier |
| Explicit restore owner assigned | Next action routes to that owner | Actual recovery still has to be tested |
Changing an assignment changes who should act, not whether the task was completed. The custom agreement retains the current roles for editing; the unmanaged preset assigns guest and application work to the owner. Reset discards those local changes and returns to the managed example. These are invented contracts with visible rules, not a reading of a provider's SLA. Shared responsibility EC2 operating responsibilities
Compare the same criteria
| Criterion | Managed service | Unmanaged service |
|---|---|---|
| Platform maintenance | Some patching, monitoring and operational work included | Customer usually operates guest software |
| Access/control | Provider may restrict risky changes | Greater access to system configuration |
| Support | Scope and response commitments can be broader | Often limited to underlying infrastructure |
| Backups | May be included but retention/restore terms vary | Must arrange unless explicitly included |
| Total cost | Fees can cover operational work | Lower sticker price can hide time and risk |
| Best fit | Limited operations capacity or standardized workloads | Special requirements with capable operators |
AWS's responsibility guidance shows why this depends on service type: a VM customer patches the guest OS, while more abstract services move some platform work to the supplier. Responsibilities for content, identity, application security and compliance do not vanish. Service-dependent duties
What to ask before paying
Read the plan, not just the marketing label. Does “managed” cover the operating system, web server, runtime, databases, CMS plugins, mail, TLS renewal, malware cleanup and restoring from backups? Is monitoring merely an alert, or does it include incident response? Can you export content and configuration? What happens when the application breaks at night?
For an unmanaged VM, verify that you can safely maintain access, install updates, rotate credentials, restrict network exposure, make backups and recover after a failed change. The freedom to administer is also an obligation. Instance security
In practice
Write a small responsibility matrix before choosing a plan. Put an owner next to operating-system updates, app updates, data backups, restore tests, DNS, TLS, monitoring and security incidents. If any row has no owner, the plan is not operationally complete. Hosting selection procedure
Common mistake: interpreting the word “managed” as an unlimited warranty. It rarely removes the need to safeguard domain credentials, application data or deployment changes.
Key takeaways
Treat managed as the beginning of a scope discussion. Separate operating-system work, application work, credentials, monitoring, backup creation and restoration. An unknown or shared responsibility needs a concrete handoff before an incident, and a named owner still needs to verify the result.
Frequently asked questions
Editorial questions, answered from the sources listed for this article.
Can backup creation and restoration have different owners?
Yes. A provider may create backups while the customer must request, authorize or perform recovery. Confirm both tasks and test the recovery path; the word managed does not settle those responsibilities.
Sources & model boundaries
Primary references used for this explanation. The interactive example is an educational model, not a live check.
- Amazon EC2 — Security
Security in the cloud responsibilities. Customers manage guest OS, credentials, updates and exposed networking on self-managed VM services.
Checked .
- Amazon Lightsail — Instances FAQ
What is a Lightsail instance?. A VPS is a virtual server with a plan-defined virtual resource bundle, not dedicated hardware.
Checked .