Redundancy in Cloudflare Workers and Implementing a Backup Server

Hey! I hope you all are doing well! I need a little bit of help or guidance regarding an issue that I’ve been facing with Cloudflare Workers. Over the past couple of weeks, I have noticed that there's been a redundancy issue with Cloudflare Workers. The service has been down multiple times, which was very unexpected. The downtime has adversely affected my application, and I am concerned about the reliability of the service. Has anyone else experienced this issue recently? On the other hand, I am looking for ways to implement a backup server that could handle the traffic when the Cloudflare Workers service is down. My goal is to have an automatic fallback mechanism to ensure that my application remains functional and accessible during such downtime. Thanks!
7 Replies
Hello, I’m Allie!
When you say "down", do you mean it was returning an error code, not resolving, showing some HTML? I haven't seen any specific Workers incidents happening recently(other than a 10 minute period of raised KV errors). Would you mind keeping an eye on when these happen, so that we can track it down?
Pato
PatoOP2y ago
by down i mean that at one point it returned only 500's but I guess that was due to issue with KV's And sure I'll keep track on this to see if it happens again!
Hello, I’m Allie!
Looks like that was just that one 10-minute period, so if you experienced it outside of that, it might be something else... https://canary.discord.com/channels/595317990191398933/895794943182909470/1116086010770632776
Chaika
Chaika2y ago
Yea you said "down mutiple times" KV was only once
Pato
PatoOP2y ago
yea was kinda of exagerating 😅
Chaika
Chaika2y ago
On the idea of redundancy though/backup, you could do something like having a route over a normal dns record pointing at a server, and you can set event.passThroughOnException() so it will hit the origin if the worker errors. Workers issues themselves are very rare though. The entire idea is Workers aren't a complicated cluster or anything like that, they are on every metal, and they run on every metal. There's no "routing to a worker from the edge", they are right there, on the edge. Workers are the fallback I would use lol
Pato
PatoOP2y ago
I see, sounds like a good idea! Thanks!

Did you find this page helpful?