We get it. When you hear "lifetime"-anything, you always think of cost-cutting on all sides. In terms of website hosting, lifetime offerings are also always connected to overcrowded servers. So, we want to take the opportunity to show you how we want to avoid all that to keep your Ghost CMS website running efficiently.
Avoiding Overcrowding
The great thing about Ghost is that it is extremely efficient. Even at peak traffic times, resources are used in a very sustainable way. This makes Magic Pages' job as a Ghost CMS hosting company pretty easy. We just need to make sure we don't overcrowd our servers.
The official Ghost documentation says that an instance of Ghost requires at least 1GB of memory. We provision well above that minimum for every customer site, with plenty of headroom for traffic spikes. Overcrowding? Not with us!
Compute Infrastructure and Data Layer
Magic Pages runs on a Docker Swarm cluster spread across multiple Hetzner data centres in Germany, with automated failover between Nuremberg and Falkenstein. If something goes wrong in one location, your site continues to serve from the other.
And your site's database? It lives on a dedicated MySQL cluster running on bare-metal servers. We use multi-primary replication so that database operations continue without downtime even during maintenance.
We also keep all our resource-intensive tasks, such as backups, maintenance runners, or updates, on completely separate servers. This separation ensures that the performance of our customers' websites is not impacted by these resource-intensive tasks.
Monitoring
We run a separate monitoring system that checks every customer site continuously and alerts us the moment something looks off, so we can step in and fix issues quickly.
And you can see this for yourself. Inside the Magic Pages customer portal, the Support tab has Website Health section with your site's real-time uptime and PageSpeed scores.
And for everything else that keeps Magic Pages running, you can check the public status page any time. It shows 90 days of uptime history for every service, plus any active incidents or scheduled maintenance.
