The Work Nobody Sees
Operating system updates, security patches, expiring certificates, disk space, log rotation, backup verification, monitoring. Continuous, invisible, and the reason most small-business compromises happen when it is not done.
The distinction between managed and unmanaged hosting is exactly this work, and it is frequently discovered rather than chosen. Buying unmanaged hosting on price, having a developer configure it once, and nobody touching it for three years produces a server accruing unpatched vulnerabilities indefinitely. Its real cost is not the monthly fee.
The Four Things That Actually Cause Outages
Rarely hardware. In practice:
The disk fills with logs and everything stops. Unglamorous and common.
A certificate expires and browsers refuse the site. Always at a weekend.
An unattended update breaks something, or a deferred one leaves a hole open.
A manual change made months ago is silently undone by a deployment, or survives and means no environment matches production any more.
All four are prevented by monitoring and discipline rather than by better hardware, which is why operational practice matters more than specification.
What Should Be Watched
The honest test is whether you find out about problems before your customers tell you.
Uptime from outside, so you know it is reachable. Error rates, with an alert when they rise. Disk space. Certificate expiry. Whether backups actually completed, since silent failure is normal.
None of that is expensive and most is free at the scale most businesses run. Its absence means your monitoring is your customers.
What Moves The Price
Managed or unmanaged, which is really the question of who does the work above.
Number of environments. Production alone, or staging and development too.
Response expectations, particularly out of hours.
How reproducible the setup is. A server configured by hand years ago is more expensive to support than one defined in code, because every change carries risk.
Compliance requirements, which add evidence and retention obligations.
Where This Goes Wrong
Nobody owns it. The most common state. The developer who set it up has moved on, the hosting company assumes you are managing it, and you assume they are.
Access never reviewed. Former contractors and old keys accumulate. Nobody removes them because nobody is looking.
Backups that have never been restored from. The most common false sense of security in small business IT.
Staging exposed to the internet with a copy of production data, protected less carefully than production. Attackers look for this specifically.
Questions To Ask Whoever Supports It
- Is this managed or unmanaged, and who patches it? If the answer is nobody, that is the finding.
- What is reachable from the internet?
- When did we last verify a restore? Not whether backups run.
- Who has access, and when was that last reviewed?
- Could we rebuild this if the server was lost? And how long would it take.
A Brief You Can Send Anyone
What we run, and where: [systems and hosting provider]
Who set it up: [and whether they are still involved]
What is monitored now: [including “nothing”]
What downtime costs: [per hour, honestly]
Hours we need cover: [business, extended, or genuinely 24/7]
Compliance obligations: [if any]
Anything we know is wrong:
How We Work
Establish what exists and what is exposed first, because that usually finds something. Then monitoring, patching on a cadence, verified restores, and access reviewed twice a year.
Where the systems themselves need ongoing attention rather than only the infrastructure beneath them, that sits in a support retainer alongside this.
Get in touch and we will start with what is currently reachable from the internet.