How sites fail in practice
In practice, most outages don’t look like “server not responding”. They look like this:
- SSL certificate expired. The site works, but the browser shows a full‑screen warning and most visitors close the tab. This happens when auto‑renewal stops and no one checks it.
- Domain expired. The renewal reminder went to spam or an ex‑employee’s address. The site and email stop working at the same time.
- Plugin update caused a critical error. The homepage shows a white screen or “A critical error occurred on this site”. It affects only some pages, so a quick glance at the homepage reveals nothing.
- Form not sending. The site looks normal, the form shows “sent”, but messages never arrive because the host changed mail settings. Such outages last months with no symptoms.
- Site works but loads in 15 seconds. Overloaded shared server or bot attack. Visitors consider the site down.
- Site has been swapped or redirects after a hack, often only for some visitors.
Google’s documentation explains how it treats sites returning 5xx errors or timeouts: during a prolonged outage it reduces crawling, and sites unavailable for a long time can drop out of the index. An outage nobody knows about for a week costs more than a week of customers.
What monitoring should check
| What | How | What it detects |
|---|---|---|
| Availability (HTTP) | Request to the homepage and several subpages every few minutes, expecting a 200 code | Server outage, critical error, bad redirect |
| Content | Check that the response contains specific text (company name in the footer) | White screen with 200 code, site replacement, hosting page "account suspended" |
| SSL certificate | Expiration date, chain validity | Expiration in X days, configuration error after server change |
| Domain | Expiration date in the registry, DNS records | Unrenewed domain, DNS change without the owner's knowledge |
| Response time | Measurement from multiple locations | Server overload, slow database |
| Form | Test send once a day and check if the message arrived | Silent email outage |
| Spam keywords | Content scan for typical breach keywords | Spammy subpages, injected links |
| Blacklists | Check domain and IP in Google Safe Browsing and anti‑spam lists | Google warning, email block |
Self‑managed monitoring
Basic availability monitoring can be set up yourself in half an hour. Services like UptimeRobot, Better Stack or StatusCake have free plans with checks every 5‑10 minutes and email notifications. That's enough to know the server is not responding.
What free plans usually don’t do: content checking (a white screen with a 200 code looks to them like a working site), form testing, domain monitoring, breach scanning. And they don’t do the most important thing: they don’t fix anything. A 3 am notification that the site is down is only useful if someone knows what to do with it.
Checklist for setting it up yourself:
- HTTP monitor for the homepage and 2‑3 key pages (offer, contact, cart).
- Check for a keyword in the response, e.g., the company name from the footer.
- SSL certificate monitor with a 14‑day before‑expiration alert.
- Calendar reminder to renew the domain, one month before the deadline, sent to two people.
- Email, SMS or messenger notifications to the person who has hosting access.
- Backup of the site off‑server, because monitoring detects an outage but doesn’t restore the site.
How much an hour of a down site costs
A simple calculation worth doing for your own site before deciding if monitoring is needed. A service‑company website generates, say, 20 form inquiries per month, of which 5 turn into jobs worth C$2,449. That means one inquiry is on average worth C$619. Inquiries mainly come on business days between 8 am‑6 pm, about 220 hours a month. A form outage that goes unnoticed for 10 days equals roughly 7 lost inquiries, or about C$4,249. A full‑site outage over a weekend is less, but a warning about an expired certificate for a week during an ad campaign costs the campaign budget spent on traffic that fled from the red screen.
This is a sample calculation with assumed numbers, not a statistic. Plug in your own: how many inquiries you get, how much one is worth, and how long past outages lasted before anyone noticed. If the result exceeds the annual monitoring cost, the decision is clear. If the site is just a brochure page without a form or ads, a free availability monitor and a calendar domain reminder are enough.
Monitoring vs Care: what to choose
Monitoring tells you something is broken. Care means someone fixes it and also does things that reduce failures: updates plugins in a controlled way, keeps backups off‑site, renews the certificate before it expires. If you have someone in your company who knows hosting and WordPress, monitoring alone is enough. If not, monitoring without care only gives an early warning about a problem you still don’t know how to handle. In Care packages monitoring is included, so you don’t have to buy both separately.
When to order monitoring
When there’s no one in your company to respond to an alert, or when the site generates revenue: an online store, a site with an inquiry form, a landing page with an ad campaign. The cost of an hour of a non‑functioning store during a campaign is usually higher than the monthly cost of monitoring with response.
At Proofsite, Guardian monitoring costs C$69/month, price list on the monitoring page. We check your site every 30 minutes: availability, content, certificate, domain, response time, blacklists. When something stops working, you receive a message describing exactly what happened and what needs to be done. One‑time monitoring setup for complex infrastructure (multiple servers, API) is Guardian Setup starting at C$369. If you want someone not only to write but also to fix, restore backups and update plugins, our site care packages start at C$119/month, with monitoring included.
Before choosing monitoring, whether in‑house or outsourced, it’s worth knowing the condition of your site. Website audit for C$369 checks over 60 items, including what monitoring will later watch.
Frequently asked questions
How to check if my site works for others, since it opens for me?
Open the site in your browser’s private mode and from a phone on cellular data, not on corporate Wi‑Fi. The browser caches a copy, so the owner’s site can "appear" to work for hours after an outage. An external monitor checks the site from other networks and avoids this problem.
How often should monitoring check the site?
For a business site, every 15–30 minutes is enough. For an e‑commerce site during an ad campaign, every 1–5 minutes. More frequent checks only make sense if someone can react immediately.
Is free monitoring enough?
It can detect a non‑responsive server, yes. It won’t catch a white screen with a 200 code, a broken form, an expiring domain, or a hack. And it won’t fix the issue, just alert you.
What happens when monitoring detects an outage?
With the DIY option you get a notification and handle it yourself. With us you receive a message describing the cause and what to do. In our Care packages we fix it: restore a backup, roll back an update, renew the certificate.
How much does website monitoring cost?
Free monitoring plans check availability every 5–10 minutes. Our monitoring every 30 minutes with content, certificate, domain and blacklist checks costs C$69/month, no fixed‑term contract.
Sources of prices and data
- Google Search Central, how Google treats HTTP and network errors (5xx, timeouts)
- web.dev, basic web metrics (Core Web Vitals)
- Google Search Central, "Security Issues" report in Search Console
Market prices are indicative data from the listed sources, checked 8.10.2026. Our prices come from the current cansitelab.pro price list and are final.
Monitoring every 30 minutes for C$69/month →