Pagecord downtime today - Cloudflare incident
Lots (most? all?) of Pagecord customers have been affected by a major incident with Cloudflare today. Sorry about this! It's very frustrating, I know.
The Pagecord servers are actually humming along just fine, and you can still post by email (like this post!), but unfortunately Cloudflare sits in between your browser and our servers so it's preventing sites being accessed over the web.
Pagecord uses Cloudflare in a number of ways:
- As a web proxy to speed up web page delivery across the world
- DDoS protection
- For storing images in their R2 object storage system
- For optimising images delivered to browsers for better site performance
- DNS
If Cloudflare has an issue in any of these areas, Pagecord will be affected. Most of these could be worked around during an incident, with the exception of object storage, but unfortunately having DNS also handled by Cloudflare prevented any remedial action from being taken today.
It's worth noting that this is a highly unusual outage for Cloudflare who are usually pretty rock solid. As of writing this, the incident has been ongoing for a couple of hours:
Pagecord operates on a shoestring and Cloudflare are very generous with their free tier, so the benefit of using them is still advantageous despite the risk of rare outages like this. It's arguably just as likely that our own compute provider (Hezner) has issues like this, so it's hard to blame Cloudflare. You can't really avoid outages over time – nothing is 100% rock solid, most certainly not on our budget.
I will definitely re-evaluate where Pagecord's DNS is managed (suggestions welcomed), but for the time being, and given the lack of budget available to Pagecord, it makes sense to keep faith in the Cloudflare service and take outages like this on the chin.