Website not responding? Before looking for a security breach or calling your developer, check whether the outage comes from the host. The status of six web hosts of the French market is read every five minutes from their official status pages; for o2switch and PlanetHoster, which have none, the page says where to look.
Page maintained by Léo Guillaume, freelance developer near Rennes, France, who has been bringing websites and software hosted with these providers back online since 2019.
Updated on
You or the host?
When the host reports nothing, the outage is almost always on the site, the domain or the cache. Five minutes are enough to find out.
If another site hosted in the same place responds, the outage does not come from the host but from your site, your domain name or your cache.
“This site can’t be reached” or a timeout point to the server or the DNS; a 500 error or a blank page point to the site itself (PHP, plugin, database).
An expired domain or DNS records changed the day before give the same symptoms as an outage. The host’s customer area or the registrar shows the expiry date.
What you need to know
Facts re-read on 27 October 2026 on the pages of each host.
6 hosts out of 8
OVHcloud, IONOS, Hostinger and Gandi use the Statuspage service; Infomaniak and LWS run their own page. That is where an incident is acknowledged, dated, then closed, and that is what this page reads.
Source: OVHcloud, Web Cloud status pageo2switch and PlanetHoster
o2switch publishes its maintenance and incident announcements on its X account; PlanetHoster closed the forum that served as a status page, and its links now lead to a contact form. For those two, this page shows no status: it points to their public announcements and their support rather than guessing.
Source: o2switch on X21 to 23 October 2026
For three days, the hosts rolled out the PHP 8.5.12, 8.4.27, 8.3.36 and 8.2.35 fixes server by server, with short interruptions announced on their status pages and X accounts. The sites that showed a blank page that morning were not hacked: they were waiting for their server.
Source: IONOS, status pageCode 200
The response code of a status page says nothing about the hosting: only its content matters. This page therefore reads the declared incidents and the affected components, never the mere fact that the page loads. And when the read fails, it says so instead of showing “all good”.
Source: Statuspage, public APIA reading, not a diagnosis: this page says what the host announces. It does not test your site. To find out why your site does not respond, send me its link: I take a look and reply within 24 hours.
During the outage
In order, without breaking anything.
Open your site from another network (4G rather than your router) and in private browsing. If it works elsewhere, the problem is local: cache, router DNS, firewall.
An incident announced by the host gets fixed without you: no need to touch the site. Note the start time for your own communication.
A word on your Google listing, your social networks or your voicemail avoids the phone calls. Give the return time announced by the host rather than a promise.
One isolated outage is no reason to change host. Three outages in a quarter, or a support team that stays silent, are: a migration takes a week to prepare.
Frequently asked questions
Nine times out of ten, the outage is on the site itself and not on the server: a WordPress plugin broken after an update, a PHP file in error, a saturated database, an expired HTTPS certificate or a domain name that has lapsed. The host only sees its machines; your site also depends on your code and your domain. The three checks on this page tell the difference in five minutes.
A program reads the official status pages of OVHcloud, IONOS, Hostinger, Gandi, Infomaniak and LWS every five minutes and keeps the time of each change of state. o2switch and PlanetHoster have no readable status page: their card says so and points to their announcements on X and their support, without guessing anything. Nothing is guessed from the response code of a page: a status page answers “200 OK” even in the middle of an outage.
Most incidents published on status pages last from twenty minutes to three hours. Long outages, from several hours to several days, are rare and usually involve failing hardware (a file server, a cluster) or a whole data centre. Hosts rarely give a return time before they have identified the cause: rely on the updates of their page rather than on an estimate.
It depends on the specific terms of your plan. The shared hosting plans of the big hosts usually include an availability commitment and a goodwill gesture on request, rarely an automatic refund. Keep the start and end time of the incident as shown on the status page: that is the evidence you will be asked for.
Not for an isolated outage: every host on this page had incidents in 2026. Change if the outages repeat, if support does not answer during the incident, or if your site has become too important for shared hosting at a few euros a month. A well-prepared migration takes a week, with no visible downtime, changing the DNS last.
No. Dibodev is an independent developer who puts websites and software online with these hosts, and maintains them, for small and medium businesses. This page exists because the first question of a client whose site stops responding is always the same: “is it me or the host?”. For an incident on your account, your host’s support remains the only contact.
Blog
When the outage comes from the site and not from the host.
Two critical vulnerabilities in PHP were patched on October 21, 2026. Are your WordPress sites, PrestaShop or WooCommerce stores, custom applications, or internal networks at risk? Learn who is impacted, what your hosting provider is doing, and what actions you should take today.
Since October 18, 2026, stolen UpdraftCentral keys have been used to install fake plugins and hidden accounts on WordPress sites—even those kept up to date. Here are the seven checks to perform today, in the exact order I follow them on a client’s site.
Restoring a backup isn’t enough: if done too late, it may still contain the backdoor; if restored without patching the vulnerability, the site will likely be compromised again within days. Here’s how to select the right backup, restore it properly, and seal the breach immediately afterward.
OVHcloud web hosting: Where to find your PHP version, how to switch to 8.3 or 8.4 from the customer portal or the .ovhconfig file, which version to select, and what OVH announced on October 21st.
At o2switch, the default cPanel PHP version is 8.1, which no longer receives official patches. Where to navigate to switch to 8.3 or 8.4, which version to choose, and what o2switch did on October 21st.
Send me the link to your site: I find out where the outage comes from (domain, certificate, PHP, plugin, database) and tell you what to do within 24 hours. Free, no commitment.