35 Replies
Project ID:
3160979a-66a1-4ab0-b77f-1ef3fe72acf43160979a-66a1-4ab0-b77f-1ef3fe72acf4
Accessing from NL in EU
#🚨｜incidents
Through VPN in US it works
But unfortunately our target audience is NL
@Pepijn - we are working on recovering the workloads as we speak. We apologize for the downtime, we are mitigating hostile traffic to on our EU proxies.
Any ETA on resolving the issue? Last update is almost an hour ago, and running ecommerce backend that is nog down for already a few hours...
None, but we are seeing some proxies restore, at times. We are updating the #🚨｜incidents with more info on where we are at.
@Pepijn, we are seeing synthetics recover, can you confirm if there is any recovery?
Still down
Should we start migrating services or is there any resolution in sight?
If you can migrate and if it will recover your service, go ahead, we want whats best for your business.
Is there any way to re-route the european traffic through the US, using Cloudflare?
fyi I got our nextjs app migrated to Vercel in about 15 minutes, still connecting to railway for the database
Cloudflare
Tunnel | Zero Trust App Connector | Cloudflare
Connect applications, servers, and other resources to Cloudflare's network via encrypted outbound tunnels. No publicly routable IPs or VMs required.
was wondering that myself - was going to experiment with a cloudflare worker, but I see that my site is working again atm. If it breaks again, I'll probably mess around with that.
@sohcah - did it restore for you?
It was back briefly - not working again now.
Yep, we will mitigate and then a new flood of IP requests.
Still working on it.
😦
Wishing you all the best in getting it resolved!
Again, terribly sorry.
I am sure you guys are doing your best so just want to say thanks and hope it's resolved soon 😦
keep us in the loop, seems interesting solution in any case 🙂
Just tried the tunnel approach myself. I ran up a new service with Docker Image:
cloudflare/cloudflared
Then set the Start Command to cloudflared tunnel --no-autoupdate run --token <mytoken>
Then I could just reference services with their private network names.sohcah's approach to this is preferred over running cloudflared inside of the same container, as its a more robust method for lack of a better word
My main API is now working through a tunnel 🙂 Not going to bother moving everything to the tunnels, but planning to move the main website too.
Thank you @Pepijn for the suggestion to use Cloudflare Tunnels 🙂
Yes, better solution @sohcah. Good workaround for now
Folks thank you- going to evangelize this mitigation, we were unsure if we could have bypassed the proxies that way.
After two of our network engineers came online, we are working on systematizing it.
Some of my services without tunnels seem to be online again. Not sure how long that will last - but here's hoping 🙂
Is tunneling actually bypassing egress fees as well?
Or is all the egress in the tunnel instance
you'd be charged for the egress of the tunnel
Yeah exactly
Well good to have this resolution in the pocket for the future, but hopefully it won't be necessary 🙂
theres now a template for it! https://railway.app/template/cf-tunnel
I'm having a hard one getting this right, wouldn't I be charged egress fees anyway ? What's the difference if it's for the tunnel to cloudflare instead of the 'outgoing connection to user x' ? Or am I just dumb ? 👀
yeah youre charged for egress still, tunnel or not, egress is egress
Well, looks like I might not need to keep migrating more services to use tunnels 😄
yeah traffic is recovering #🚨｜incidents
incident has been mitigated, closing out 