It tells you the server was up when you looked. It says nothing about whether the certificate behind the padlock expires next month, whether the backups have ever been restored, whether anyone would notice if the site went down at 2am, or whether the domain renewal notice is going to an email address belonging to a developer who left two years ago.
The businesses that get caught out are rarely running anything exotic. They have a website, maybe a booking system or an internal tool, a domain, and a server somewhere. What they usually do not have is anyone whose job is to watch those things. Often the infrastructure was set up well by someone capable, and then that person stopped being available. Everything still works, until something with a deadline attached reaches that deadline.
The five areas below are the ones that most often sit unmaintained in small businesses. For each one there is what it actually is, how to check it yourself, one concrete thing to do this week, and the point at which it becomes a job for someone technical rather than an afternoon of your own time.
1. SSL/TLS certificates
What this is. The padlock in your browser's address bar means your site presented a certificate, and the browser checked it and accepted it. That certificate is issued for a fixed period, and it has an expiry date on it. When it expires, browsers stop accepting it and show visitors a full-page warning instead of your site. Nothing is broken on your server. The certificate has simply run out.
Certificates used to be renewed annually and often by hand. Now they are commonly renewed automatically, but the intervals are shorter than people expect. Let's Encrypt, the free certificate authority used by a very large share of the web, issues certificates valid for 90 days and recommends renewing them every 60 days. The industry is moving to shorter lifetimes still: the CA/Browser Forum, which sets the rules certificate authorities follow, has voted to reduce the maximum certificate lifetime from 398 days to 47 days over the coming years. Renewal is becoming a routine, frequent, automated job, not an annual errand.
How to check it. Click the padlock (or the icon to the left of the address) and open the certificate details. You will see the issuer and a "valid until" date. Repeat this on the exact addresses your customers use, which is more than one: yourdomain.com, www.yourdomain.com, and any subdomain like booking.yourdomain.com or portal.yourdomain.com can each have their own certificate with its own expiry date. Check on a phone using mobile data rather than office wifi, and check from outside the office network if you can. People inside a network sometimes see an internal page or a cached one and never see what customers see.
One thing to do this week. Find the expiry date of your main site's certificate and put it in a calendar with a reminder four weeks before it. Then ask one question of whoever manages your hosting: is this certificate renewed automatically, and who gets notified if the renewal fails? If the answer is a shrug, you have found something worth fixing.
When to bring in someone technical. When renewal is not automatic and nobody owns it. When you have several services, subdomains or a wildcard certificate. When a renewal has already failed once, since that usually means the automated validation is misconfigured and will fail again. When the certificate covers the wrong name, or when visitors report warnings you cannot reproduce.
2. Backups
What this is, and what it is not. A backup is a copy of your data kept somewhere it can survive whatever happens to the original. Having one is not the same as knowing it works. A backup job can report success every night for a year and still fail when a restore is attempted. The reason is often something nobody checked:
- The backups are written to the same server, or the same provider account, as the live data. One failure takes both.
- The job quietly stopped months ago after a password change or a full disk, and no one noticed, because nothing alerts on it.
- The backups are encrypted, and the passphrase lives in the head of the person who set them up.
- The database is copied while it is in use, so what gets saved is not a clean, restorable database.
- The one copy exists, but it is in the same building.
The widely used mental model is the 3-2-1 rule: three copies of the data, on two different kinds of media, with one copy off-site. The point of the off-site copy is that it survives fire, theft, ransomware, and a hosting provider's mistake.
How to check it. Ask where the backups go and who set them up. Then ask whether anyone has ever restored one. That second question is the one that matters, and a long pause is a real answer. Open the backup dashboard or the log emails and confirm there is a recent, successful run, that the destination is not the same machine as the live data, and that someone other than one person knows it exists.
One thing to do this week. Restore something small and low-stakes into a scratch location: one folder of documents, or a test copy of the database. Do it with whoever knows the system, and write down the steps as you go. A restore you have actually performed once is worth more than any backup dashboard.
When to bring in someone technical. When nobody can say how the data would come back. When the encryption key is missing or unknown. When there is no off-site copy. When only one person knows how any of it works, particularly if that person is a contractor you no longer have a relationship with. When you need to answer two business questions in writing: how long could you be offline before it costs you real money, and how much recent data could you afford to lose.
3. Uptime and service monitoring
What this is. "It's working" means it worked the last time a human looked. Monitoring means something checks your service from outside, automatically at regular intervals, and tells a person when it stops answering. Without it, your monitoring is your customers, and they will tell you by leaving rather than complaining.
A useful distinction: you want to be told when your site does not load, not when a server's memory usage crossed an arbitrary line. Healthy monitoring watches what your customers experience, and treats everything else as detail.
How to check it. Ask the question directly: if our website went down right now, how would we find out, and how long would it take? Then trace the chain. Monitoring that covers only the web page misses the parts that fail more often: the domain registration, the DNS records that translate your name into a server address, the certificate, the disk filling up on the server, and background jobs that never ran.
One thing to do this week. Set up one external uptime check, pointed at your main web address, that notifies a real person by email or SMS when it fails. Free and low-cost tiers exist for exactly this and take minutes to configure. Make sure it alerts someone who will actually look at their phone.
When to bring in someone technical. When alerts already exist but go to an address nobody monitors, or to a former employee. When there are so many alerts that everyone has learned to ignore them; at that point alerting is worse than none, because it creates the belief that someone is watching. When you need coverage beyond the website: server disk and services, backup jobs, certificate expiry, DNS changes, and internal tools that only your staff use.
4. Domains and DNS
What this is. Your right to use a domain name depends on keeping its registration active. It is registered for a period, usually a year, and it has to be renewed with the registrar it was registered through. Your DNS records live separately from the domain registration, sometimes at the same company and sometimes not, and they are what actually point your name at your website, your email, and anything else customers reach at your address. Two things follow from that.
First, renewal notices go to a contact address on the registrar account. If the domain was registered by a web developer or an agency, that address may be theirs. The notices still arrive, just not to you. Second, missing the renewal does not make the name immediately available to anyone else. ICANN rules require registrars to remind the registrant about a month and again about a week before the registration expires. If the name is not renewed, the registration is eventually deleted, and the domain then enters a 30-day Redemption Grace Period during which DNS resolution is disabled but the registrant can still restore it, usually for a fee. Only after that period ends does the name become available for someone else to register, and recovery is no longer guaranteed.
How to check it. Use ICANN's own Lookup tool at lookup.icann.org to see which registrar holds your domain and what status it is in. Then log into that registrar's account and confirm you have working credentials, that automatic renewal is switched on, that the expiry date is what you expect, and that the contact email is an address a current person reads. Next, export or photograph the DNS records so there is a written record of what each one does. That record is what saves you the day someone needs to change hosting and nobody knows which entry points at what.
One thing to do this week. Log in, turn on auto-renewal, and set the account contact email to an address you control. Then save a copy of the DNS records somewhere more than one person can find. If you cannot log in at all, that is the finding, and it is better to discover it now than on the day your site goes dark.
When to bring in someone technical. When you cannot get into the account, or the domain is registered in someone else's name. When you are changing hosting, email providers, or a website platform. When DNS changes are needed for anything with mail records attached, where a small mistake can silently break your email. When nobody can explain what half the records are for.
5. Who receives the alert, and who acts on it
What this is. This is the part that decides whether any of the above actually helps you. An alert that lands in a shared inbox nobody reads, an address belonging to someone who left, or a support queue at a hosting company is not monitoring. It is a log line. Someone has to receive the alert, recognize what it means, and be able to act, ideally without first working out who has the password.
How to check it. Ask what happens at 3am on a Saturday if your site goes down. Then look at where your alerts are configured to go, and search that inbox for the last alert it received. Did anyone respond? Who? This question exposes more unmaintained infrastructure than any technical check, because it is possible to have monitoring, backups and certificates all correctly configured and still have nobody responsible.
One thing to do this week. Write one page and keep it where more than one person can find it: what is monitored, who receives each alert, who is next if they do not answer, and what the first action is. Names, not roles. If the page has one name on it, you have found your single point of failure.
When to bring in someone technical. When one person holds all the access and the knowledge. When you need cover outside working hours and there is nobody to provide it. When you want the responsibility for watching all of it transferred off your own plate, so that finding out something broke is not something that depends on you happening to check.
A five-minute self-check
- Click the padlock on your site and note the certificate's expiry date. Do the same on
wwwand any subdomain. - Ask who can restore a backup, and when one was last restored.
- Ask how you would find out about an outage at night.
- Log into your domain registrar and confirm auto-renewal, expiry date, and that the contact email is current.
- Search your alert inbox for the last alert and check whether anyone acted on it.
- Write down the names of the people who currently hold access, and count them. If the number is one, that is worth thinking about.
None of these are difficult. They are also the things that go unexamined for years, because they sit quietly in the background working, right up until something with a deadline arrives.
IronRelay Systems helps small businesses review and maintain this infrastructure when there is no one internally responsible for it. If you want a clearer picture of where your systems stand, the Remote Infrastructure Audit provides a fixed-fee review and a prioritized list of what needs attention.