Network maintenance in Amsterdam on 14 October

On 14 October we are upgrading the core routers in our Amsterdam location. The work is planned so that traffic moves to a backup path before anything is touched, and most servers will not notice it at all. This post covers the timing, what to expect and what you can do to be ready.

When it happens

We picked the quietest hour of the week based on traffic across the Amsterdam network.

What we are doing

The new routers give us more capacity toward our upstream carriers and internet exchanges, which means more headroom during busy evenings and even better absorption of large DDoS attacks. The upgrade happens in stages:

1. Traffic is moved from the first core router to the failover path.
2. The first router is upgraded and tested.
3. Traffic moves back, and the same steps repeat for the second router.

Each switch between paths is the moment where a short blip can happen. In practice this means a few seconds of loss, twice, inside the hour.

What you might notice

Most services will see nothing. Connections that are sensitive to even a short interruption may react:

• SSH sessions can freeze for a moment and then continue.
• Game servers may show a brief lag spike; players should stay connected.
• Long running transfers might pause and resume.
• Monitoring tools with very short timeouts may log a single failed check.

Your servers stay powered on the whole time. Nothing is restarted, and no data is touched.

Tip: If you run an uptime monitor with alerts, pause it or raise the timeout for 03:00-04:00 CET on 14 October, so you do not wake up to a false alarm.

Do you need to do anything?

Usually not. If you want to be extra careful:

• Avoid scheduling backups or large transfers to remote storage during the window.
• If you run a database cluster across locations, make sure failover timeouts are longer than a few seconds.
• Tell your game community in advance, if they are likely to be online at 3 in the morning.

Following the maintenance

We will post updates on our status page when the work starts, after each stage, and when it is finished. If something does not go to plan, we roll back to the current setup, which is ready and tested.

If your server shows problems after 04:00 CET that it did not have before, run a quick check from the service page: look at the Overview and Network tabs, and open the VNC console in the Controls card if the server does not respond. Then go to Support and press Open ticket, mentioning this maintenance, and our network team will look at it first.

Thanks for your patience while we make the Amsterdam network faster and more resilient.