Most small business outages are discovered by a customer, hours after they begin, because nobody was watching. Monitoring is inexpensive and the important part is not the alert but who receives it and whether that person can act. A provider who tells you before you notice is buying you the only thing that matters during an outage, which is time.
How outages are usually found
The typical sequence is not a monitoring alert. It is a customer mentioning it, or the owner checking something on a Sunday, or an email bouncing. By that point the site has often been unreachable for hours, sometimes since overnight.
Nobody is at fault in the ordinary sense. Nobody was assigned to watch, so nobody watched. The site sits in a state where it either works or it does not, and the business finds out through the same channel as everybody else.
The cost is invisible, which is the whole difficulty. You are never told about the person who searched, clicked, got an error, and called the next company on the list. There is no notification for that, and it is the largest part of what an outage takes from you.
What actually goes wrong
- The server itself. Hardware failure, a network problem at the data centre, or a neighbouring account consuming resources on shared infrastructure.
- An expired certificate. The site is up and every browser refuses to show it, which reads to a visitor as a security warning about your business.
- A domain or DNS failure. Nothing wrong with the site at all, and it is unreachable everywhere. The renewal failure version of this is entirely preventable and still common.
- An update that broke something. A plugin, a platform version, a PHP upgrade applied by the host on a schedule you were not told about.
- Traffic you did not expect. A busy season, a mention somewhere, or a bot. The site does not fail so much as slow to the point of being useless.
- A compromise. The site is technically up and serving something other than what you built, which is worse than being down.
These fail differently, and a monitor that only checks whether the server answers will happily report everything as fine during several of them.
What monitoring should actually cover
Whether the page loads, not whether the server responds
A server can return a response containing a database error. Useful monitoring checks that the page renders and that expected content is present on it, which catches the failures a simple ping reports as healthy.
Certificate expiry, in advance
Checked weeks ahead rather than discovered on the morning it lapses. Automated renewal usually works and occasionally does not, and the failure mode is a browser warning shown to every visitor.
Whether forms still deliver
The most commonly missed failure by a wide margin. The site is up, the form submits, the confirmation appears, and the mail goes nowhere. Businesses have lost months of enquiries this way without a single symptom visible from the outside.
Speed, tracked over time
Degradation is more common than outright failure and considerably harder to notice, because it arrives gradually and you adapt to it. A site that has slowly become twice as slow is losing enquiries continuously rather than in a single visible event.
Checks from more than one location
Regional network problems are real. A single check point produces both false alarms and missed outages, and neither is useful.
The part that matters more than the monitoring
An alert is only worth what the response behind it is worth. Most businesses that do set up monitoring send the alert to the owner, who receives it at 2am, cannot do anything about it, and forwards it in the morning to someone who then begins investigating.
The useful arrangement sends the alert to whoever can actually fix it, with the access already in place to do so. That is the substantive difference between hosting where you are the escalation path and hosting where somebody else is, and it is most of what people are describing when they talk about managed hosting.
The question worth asking a provider is not whether they monitor. Almost everyone says yes. It is what happens at 2am on a Sunday: who is notified, who acts, and whether you are told before you notice.
What silence costs
Work it out with your own numbers rather than a general figure. Take your monthly enquiries, divide down to a daily rate, and multiply by what an enquiry is worth to you. For most service businesses a single day offline is a number that comfortably exceeds a year of proper hosting.
Then add the parts that do not appear in that arithmetic. Search engines that crawl during a prolonged outage record the failure. A customer who arrives at a security warning does not usually come back to check whether it has been resolved. Reputational damage from an outage during your busiest week is disproportionate to its length.
What to do this week
- Submit your own contact form and confirm the message arrives. Then check the spam folder of the receiving address, because that is where these quietly end up.
- Find out who is notified if the site goes down tonight, and whether that person can act or only relay.
- Check the certificate expiry date and whether renewal is automatic.
- Ask your host how you would be told about a planned maintenance window, and whether you have been told about past ones.
Four checks, half an hour, and they surface the most common silent failures. The form test in particular is worth doing today, because the businesses that discover a broken form are almost always discovering it long after it broke.
Frequently asked questions
How do I know if my website is down right now?
Check from a phone on mobile data rather than your own connection, since local caching and DNS can show you a working copy of a site nobody else can reach. Better still, have monitoring that checks from several locations and tells you first.
What should website monitoring check?
That the page renders rather than just that the server responds, certificate expiry in advance, whether form submissions still deliver, speed tracked over time, and all of it from more than one location.
Why did nobody tell me my site was down?
Usually because no monitoring existed, or because the alert went to an address nobody watches. Many hosts monitor their own infrastructure without monitoring whether your specific site is serving pages correctly.
How much does downtime cost a small business?
Take your monthly enquiries, reduce to a daily figure, and multiply by what an enquiry is worth. For most service businesses one day offline exceeds a year of good hosting, before counting the customers who never return.
What is the most commonly missed website failure?
Forms that submit successfully and deliver nothing. The site looks perfectly healthy from every angle, and businesses have lost months of enquiries before anyone noticed the absence.
Should my host be monitoring my site or should I?
Someone with access to fix it should receive the alert, which in practice means your provider. An alert reaching only the owner at 2am creates awareness without creating a response, and awareness alone does not shorten an outage.
West Coast Media Solutions Inc. provides web design, web development, hosting, digital marketing, and business consulting to organisations across Canada, drawing on more than twenty-five years in the field.
Would you know tonight if your site went down?
We monitor the sites we host and act on the alert ourselves, so the first you hear about a problem is usually that it has been fixed.
Start a Conversation