Date:November 21, 2025
Cloudflare outage affected large parts of the internet. Here’s how we reacted at Webnorth
On Wednesday, November 18, large parts of the internet experienced outages due to a global Cloudflare failure.
A fault in their edge network triggered a wave of 5xx errors, affecting even some of Denmark’s largest websites.
As Denmark’s leading WordPress agency, Webnorth continuously invests in stable operations, high performance and maximum uptime for our clients’ WordPress solutions. We have a full-time DevOps specialist focused on monitoring, performance and preparedness. Still, we were not immune this time.
How we detected the issue
Although the error was outside of our control, it was caught by our monitoring systems before Cloudflare issued any official statements.
Our synthetic probes suddenly registered a significant number of 5xx errors marked as “Cloudflare,” while our origin servers continued running without issues. This provided an early and clear signal that the problem was not within our own infrastructure, but in an external layer.
What we’ve done since on our WordPress Setup
Following the incident, we have further strengthened our setup. Specifically, we have:
Built scripts that can automatically disable the Cloudflare proxy and bypass the edge layer
Set up incident alerts that detect Cloudflare issues before any official announcements
Prepared an operational workflow that can be activated with a single command
We are now working towards full automation, so our systems can automatically switch away from Cloudflare during major outages and ensure that our clients’ websites remain accessible.
Conclusion
Even the most widespread and well-tested technologies can fail.
However, with robust monitoring, structured DevOps practices and clear contingency plans, it is possible to minimize the impact, even when the internet itself falters.