Client Website Down? What to Do, in the Right Order
If a client website is down, what to do comes down to a fixed order: confirm it is really down and for whom, work out which layer failed (domain, DNS, SSL, server or application), contact whoever controls that layer, and tell the client before they tell you. Once the site is back, check that nothing SEO-related came back broken.
Confirm the outage before you touch anything
A site that won't load on your laptop is not yet an outage. Try it on a different network, mobile data instead of the office wifi, and with an online checker that fetches from another location. Then look at what actually comes back:
- a timeout or no response usually points at the server or the hosting
- a
500,502,503or504means something answered, but the application or a service behind it failed - a
403or a challenge page can mean a firewall is blocking you, not that the site is down - a certificate warning means the site is up but the SSL certificate is expired or wrong
DNS_PROBE_FINISHED_NXDOMAINin Chrome means the domain name isn't resolving
From a terminal, curl -I https://example.com gives you the status code and headers in one go, and dig example.com (or nslookup on Windows) shows whether DNS still returns an IP address.
Check the domain and DNS first
An expired domain takes everything down at once, website and email included. Run a WHOIS or RDAP lookup and read the expiry date and status. A status such as clientHold or redemptionPeriod means the registrar has suspended it, usually after a failed renewal payment, and the fix sits with whoever owns the registrar account.
If the domain is fine but DNS returns nothing, or a different IP from last week, find out who touched the nameservers or records. Our guide to domain expiry monitoring covers how to stop the expired domain case from ever getting this far.
Who to contact, depending on what you found
Match the call to the layer that failed:
- domain expired or on hold: the registrar account owner, then the registrar
- DNS records wrong: whoever controls the DNS zone (registrar, Cloudflare or host)
- timeouts or
5xxacross the whole site: the host's support, with the status code and the time you first saw it - errors only on some pages, or right after a deploy or plugin update: the developer behind the last change
- invalid certificate: the host, or whoever manages renewals
What to tell the client, and when
Tell them early. A first message within the hour, even if it only says what you know so far, lands far better than a perfect explanation sent after they have rung you. Say what is affected, when it started, what you have confirmed, who you have contacted and when they will hear from you next.
Don't guess at the cause in the first message: say what you have confirmed. If several people on the client side keep asking, a shared client status page saves you sending the same update five times.
After it comes back: what to check
A site restored from a backup, moved to another server or patched in a hurry can come back different. Before closing the incident, check that:
robots.txthasn't been replaced by an old or staging version withDisallow: /- key templates carry no
noindexmeta tag orX-Robots-Tagheader - the XML sitemap loads and lists roughly the same number of URLs as before
- canonical tags still point at the live domain, not a staging host
- GA4, GTM and any ad pixels still load on the main templates
Old backups are a classic source of a stray noindex, which is why how to catch a noindex in production is worth a read.
Where Deltio fits
Deltio checks each site over HTTP every 10 minutes and marks it online, offline or firewall blocked, with the status code and response time. You get an email or Slack notification when a site goes down and another when it recovers, and every site can have a public status page to share with the client. The same account watches SSL validity, domain expiry, robots.txt, the sitemap, SEO tags and marketing tags, so most of the list above is covered. Sitemap and SEO scans run on a daily cycle, so after a recovery it is worth running a manual scan rather than waiting. What Deltio can't tell you is why the server failed: for that you still need the host and its logs.
Plans start at £24 a month, with a 14-day free trial.
Frequently asked questions
- How soon should I tell a client their website is down?
- As soon as you have confirmed it is a real outage and not a local problem, ideally within the first hour. A short message with what you know, who you have contacted and when you will update them is enough. Clients are far more forgiving of an outage than of hearing about it from a customer.
- Should I restore the site from a backup straight away?
- Not until you know which layer failed. A backup won't fix an expired domain, a DNS change or a host outage, and restoring an old copy can undo recent content and bring back old settings. Restore when the problem is clearly in the site's files or database.
- What if the site is down only for some people?
- Partial outages usually come from DNS changes still propagating, a CDN edge with problems in one region, or a firewall rule blocking certain countries or networks. Ask affected users for their location and the exact error, and test from several locations before contacting the host.
- Will a short outage hurt the client's Google rankings?
- A brief outage is very unlikely to cause lasting damage. Problems start when a site returns errors for long stretches or repeatedly, because crawling slows down and pages can eventually drop out. Returning a 503 during planned maintenance is the standard way to signal a temporary problem.
- Can a monitoring tool tell me why the site went down?
- Only partly. A monitor records the status code, response time and when the problem started, which narrows down the cause. The actual reason, such as a crashed database or an exhausted hosting plan, is in the server logs and with the host.