Hostinger Outage (Oct 20, 2026): Google Cloud Took Down hPanel, Not Your Website

What happened in the Hostinger outage on October 20, 2026

On October 20, 2026, Hostinger customers lost access to the account side of the platform: the Hostinger Dashboard (hPanel), the service API, Website Builder, the Kodee AI assistant and checkout. According to Hostinger's incident page on its status site, the cause sits upstream: Google Cloud, which Hostinger uses as a third-party cloud provider for these services, first had a network outage, and Google later told Hostinger it was triggered by a power failure at one of its data centers.

The detail that matters most for site owners: Hostinger states that websites hosted on its servers were not affected and kept serving visitors. What broke was the layer you use to manage your account, not the layer that serves your pages.

On Google's side, the Google Cloud Service Health dashboard (status.cloud.google.com) shows an active incident titled as multiple services having issues in the europe-west2 region (London), opened at 06:08 UTC on October 20 — eleven minutes before Hostinger's first update. It lists AlloyDB for PostgreSQL, Apigee, BigQuery Data Transfer Service and 28 more products. Hostinger hasn't named the region, but the timing lines up.

At the time of writing, Hostinger had not published a recovery time, because Google had not given one.

Hostinger outage timeline (UTC)

Summarised from Hostinger's status updates, plus the start of Google's incident. Times are UTC; Hanoi is UTC+7, US Eastern is UTC−4 in October.

Roughly one hour passed between the first report and a named root cause. That is a reasonable time to identify an upstream dependency, and the early, explicit line that sites were up was the most useful thing Hostinger published.

Is my Hostinger website down?

Very likely not. Hostinger's updates say hosted websites kept working throughout. If your site is failing, treat it as a separate problem and check it the usual way: your own error logs, a recent deploy, DNS, or an expired certificate. Don't assume the outage explains it.

To confirm from outside your own network, load the site from a phone on mobile data or an external uptime monitor. If it responds there, visitors are being served.

What you can and can't do while hPanel is down

This is the part most outage write-ups skip. "Dashboard down, sites up" means some daily tasks are blocked and others are untouched, depending on whether they go through Hostinger's account platform or straight to your server.

Practical takeaway: if you need to push a fix today, go through SSH, SFTP or your CMS admin instead of waiting on hPanel. If something can only be done in the dashboard, such as a DNS change hosted at Hostinger or a renewal, it has to wait.

Why the websites stayed up: control plane vs data plane

Cloud and hosting platforms are usually split into two layers:

• Data plane — the servers that actually serve your site, run PHP or Node, and answer requests from visitors.
• Control plane — the systems you use to manage those servers: the dashboard, the API, billing and checkout, the site builder, support tools.

Hostinger's updates describe a control-plane outage. The control plane runs on Google Cloud, so when Google had trouble the management layer went dark, while the data plane, the servers your visitors hit, kept running. That separation is exactly what you want from a host: a failure in billing or the dashboard should never take customer sites with it, and here it didn't.

The trade-off is that you can't change anything through the control plane until it comes back. For most sites that's an inconvenience. For a business that needs to push an urgent fix, renew an expiring service, or rotate a compromised credential through the dashboard, it can be a real problem. That's why the next section matters.

Can a single data center power failure really cause this?

Yes. Hyperscalers design for failure, but power events still take capacity down. Google's own report on a previous incident is a good example: in October 2024 an electrical arc flash in a power distribution unit in one Frankfurt zone, europe-west3-c, caused a partial power loss and degraded cooling, and Google had to shut hardware down. Multiple services in that zone were impacted for about 7 hours 39 minutes, according to Google Cloud's incident report for europe-west3.

A workload that lives in one zone, or depends on one zonal resource, goes down with that zone. This time Google's dashboard flags a whole region, europe-west2, with 31 products affected. A regional incident is harder to sit out than a zonal one: running in three zones of the same region doesn't help if the region itself is degraded. Only a second region or a second provider does.

The power-failure cause comes from Hostinger relaying Google; Google's dashboard summary did not state a root cause when we checked. Its post-incident report on Google Cloud Service Health will be the authoritative account.

How to protect your site from your host's next outage

You can't stop your provider's upstream from failing. You can make sure an outage in their management layer doesn't stop you. These are the changes worth making after today, in order of value:

1. Keep DNS somewhere independent. If your domain's DNS is managed inside your host's dashboard, a dashboard outage freezes your DNS. Moving DNS to a separate provider means you can repoint traffic even when your host's panel is down.
2. Keep backups off the host. Hosting backups that can only be restored through hPanel are unavailable during exactly this kind of event. Keep a recent copy of files and database in storage you control.
3. Make sure you can deploy without the dashboard. Confirm SSH or SFTP access works and that you have the keys on hand. Test it on a normal day, not during the incident.
4. Don't make your pipeline depend on the hosting API alone. If your CI deploys through the provider's API, give it a timeout and a fallback path over SSH, so a control-plane outage fails fast instead of hanging your release.
5. Monitor from outside. An external uptime check tells you within minutes whether visitors are affected, which is the first question your clients or your boss will ask.
6. Renew early. Domains and plans renewed on the last day are exposed to any checkout outage. Renewing a couple of weeks ahead removes that risk entirely.
7. Subscribe to your host's status page. Hostinger posted clear, frequent updates on its status site; getting them by email or RSS beats refreshing social media.

A deploy path that doesn't touch the hosting dashboard or API can be as short as this:

# Deploy over SSH — works even when the host's dashboard and API are down
rsync -az --delete ./build/ user@your-server:/home/user/public_html/
ssh user@your-server 'cd ~/public_html && php artisan migrate --force'

Swap the second line for whatever your stack needs (a WordPress cache flush, a Django migrate, a service restart). The point is the route, not the command.

If you run a business site and want it set up this way — DNS, backups, monitoring and a deploy path that survives a provider's bad day — that's part of how we build and operate business websites.

Should you leave Hostinger after this outage?

Not on the strength of this incident alone. Every host depends on someone: most run on, or buy services from, a large cloud provider, and those providers have bad days too. What you want to judge is how the host behaves when it happens:

• Did customer sites stay up? According to Hostinger, yes. The data plane was isolated from the failure.
• Was communication fast and honest? First update within minutes, root cause named about an hour in, and a clear statement of what was and wasn't affected.
• What happens after? Watch for a post-incident review and whether Hostinger reduces its dependence on a single cloud for the dashboard, API and checkout. That's the real test.

If a dashboard outage of a few hours would genuinely hurt your business, the fix is less about switching hosts and more about the independence steps above: your own DNS, your own backups, and a deploy path that doesn't need anyone's control panel.

Not sure how exposed your current setup is? Tell us how your site is hosted and deployed, and we'll point out the single points of failure. Request a free quote from Netalith — no account, no commitment.