Cloudflare timeouts some times
Hi there,
I have a project with muliple services, sometime cloudlfare cant connect ot railway, I have the SSL mode is set to Full, everything is working fine, it's just that sometimes the server does not respond.
There are 8 instances active right now and it's still happening
e0d0ed17-eedc-4575-8b89-1b314af7f950
49 Replies
Project ID:
e0d0ed17-eedc-4575-8b89-1b314af7f950there's an ongoing incident right now that's likely related to your issue #🚨｜incidents
We have this issues since 2-3 days
Also the Incident is for US-WEST, but we are on EU WEST
hmm theres no been no incidents for eu-west in the last few days, what error do you get from cloudflare?
I just deployed to US East, to see if this error still shows
are there any errors in the deploy logs?
No errors at all
what kind of app is it?
Node + NestJS
It's working fine on US East I guess
but it's very weird because I have other services on that project, it's the same monorepo they are deployed in EU West and they are working fine
very odd, but no one else has reported timeout issues until today
I will leave it in US east for today, and tomorrow I will move it again to EU, working fine on US east
okay sounds good, keep me updated!
I still have the same issue
@Brody I have tried everything. I created another service with the same configuration, if this doesn't fix it, I will have to move more than 15 services to another provider
when you visit the cloudflare protected domain and get that timeout message, what would visiting the railway domain directly show you?
visiting the railway domain would not work, same as visiting my domain
it just stays on loading
that would have to timeout at some point too right? when it does, what would you see?
it would respond after sometime, I havent seen a railway error
this is starting to look a lot like another users issue, they thought it was a railway outage, but it turned out they where getting ddos'ed and their app was soft locking
right but you can still access the app through the railway provided domain
either way, we can't rule anything out
No, I cant, it's the same thing as going through my domain; the ap will respond after 30s
And then the app automatically becomes responsive again
sorry I meant that theoretically you could access it via the railway domain, and that could be where the ddos is getting through and not showing on cloudflare
I was using a easypanel server before and we had no isses at all, I moved all our services on Railway and this started happening
I have removed the railway domain
only using my domains
that technically doesn't close off access
But a ddos attack would use resources on Railway right?
did you say all 15 of your services are experiencing the same thing?
Noo, its only one
yeah, your memory is pretty high for a node/nest app right?
there are 4 instances active
replicas?
yes sorry
okay going to ask the team
I will be looking into this inside of an hour
@thomas were you able to check
friendly request, please dont ping the team #🛂｜readme #5
Hey I am sorry, but we have a possible incident we are investigating that pushed back my estimate
you are very much still on my list, could you post the service ID?
9e89d04a-cc90-42a3-ac7f-89b05dcd314e
@Brody I think you are right about the DDoS attack
just an idea, you can sometimes tell based on billing for egress
unless cloudflair is protecting you
sorry commenting without all the context yet
you can always talk directly to the railway service with host header manipulation, bypassing cloudflare
what would that solve?
I'm talking about how you could be getting ddos'ed without cloudflare noticing
what does your network metrics look like?
All the metrics look normal to me
I only noticed that on the logs
So this is a US east only issue?
Can you explain that part?
I see the service is in EU West at the moment. To be clear, are you still experiencing this in that region?
No this is happening on all regions