Server down?

Check these first

1. Is it really down?

   Open the site over mobile data, not your office Wi-Fi. Check your hosting provider’s status page. If the data centre has an outage, all you can do is wait.

2. Does the server answer at all?

   ping -c 3 your-domain.com

   No ping, no SSH: open the console in your hosting provider’s control panel, often called “VNC” or “KVM”. It shows whether the server hangs, is booting or sits in the rescue system.

3. Is the disk full?

   df -h

   One partition at 100 % and the database and web server can’t write anymore. Usually it’s logs, old backups or Docker images.

4. Which service failed?

   systemctl --failed

   journalctl -u nginx -n 50 --no-pager

   The first line shows what isn’t running, the second shows why. Replace nginx with the service that’s missing, for example apache2, mariadb or php8.2-fpm.

5. Did it run out of memory?

   dmesg -T | grep -i "out of memory"

   If the kernel killed processes because memory ran out, it shows up here. MySQL and MariaDB are the usual victims.

6. Expired certificate or domain?

   A certificate warning in the browser often just means the SSL certificate expired. If the browser can’t find the domain at all, check whether it was renewed.

7. What changed last?

   An update, a new plugin, an edited config file? The last change is the first suspect.

What not to do

• Don’t keep rebooting. It rarely helps while the cause is still there, and it wipes traces from the logs.
• Don’t delete what you don’t know. Especially not under /var/lib, that’s where the databases live.
• Don’t reinstall before the data is safe. A new server is quickly built, lost data is not.
• Don’t send passwords by email. We work with our own SSH key.

• SSH access with an SSH key provided by us
• sudo-level access
• If SSH no longer works: access to the console at your hosting provider

What we do

--diagnoseFind the cause

Logs, services, memory, disk, network. We keep digging until it’s clear why it went down.

--restoreBring it back online

Repair services, free up space, fix the configuration. Your data stays where it is.

--reportReport

What happened, what we did and how to prevent it next time. Written into the Trello card.

--preventSo it doesn’t happen again

Monitoring, alerts for a filling disk, automatic updates. As a separate task or in the subscription.

Urgent? Express.

How it works

1. Book the task

   Pay €189 with Stripe. No account, no call, no quote.

2. Give access

   You get a Trello board and our public SSH key. In the card you write: which server, since when, what doesn’t work, what you already tried.

3. Back online

   The card shows what’s happening. At the end, the report is there.

Questions