“Fully managed” can mean anything from “we will reboot it if you ask” to “we run it so you never have to think about it.” The only useful definition is a clear list of what you no longer have to do yourself, and a clear line showing where your responsibility ends and ours begins.
The spectrum, from unmanaged to fully managed
“Managed” is a range, not a single product. Knowing where an offer sits on this range is the difference between a fair comparison and an unexpected gap in coverage.
| Level | What the provider does | What you still own |
|---|---|---|
| Unmanaged | Delivers working hardware and network | OS, patching, monitoring, backups, everything above |
| Managed OS | Patching, monitoring, backups, hardening | Your applications and data |
| Fully managed | The above plus SLA-backed incident response | Your application logic and business decisions |
Who is responsible for what
The clearest way to read any managed offer is as a responsibility matrix — the same idea as the cloud’s shared-responsibility model, made explicit. For fully managed bare metal it typically looks like this:
| Area | Provider | You |
|---|---|---|
| Hardware, power, network | Yes | — |
| OS patching and hardening | Yes | — |
| Monitoring and alerting | Yes (24/7) | — |
| Backups and restore testing | Yes | Define retention and RPO |
| Incident response | Yes (to SLA) | Be reachable for decisions |
| Application code and data | — | Yes |
What you stop worrying about
- Watching dashboards overnight — we monitor and alert around the clock.
- Chasing security patches and OS updates — handled on a defined schedule.
- Wondering whether backups actually work — they are taken and their restores are tested.
- Improvising during an incident — there is an SLA and a defined response path.
The detail that matters
Technical detail
Managed operations at Metal on Cloud cover 24/7 monitoring and alerting, OS patching and security hardening, backup management with retention, disaster-recovery testing, and SLA-backed incident response with defined response times (P1 under 15 minutes). What matters is not the length of the checklist but that the work is contractually someone else’s responsibility, with a response time you can hold them to.
Key takeaway
If you do not have a dedicated operations team, “managed” is usually cheaper than the headcount it replaces — and it removes the after-hours escalation. Confirm where on the spectrum an offer sits before comparing prices.
Ready to talk specifics?
Get a Quote