Random CPU spike and unresponsive server
We switched from Vercel to Railway one week ago, and everything was going well until today.
For some reason the CPU usage spiked (no more users than usual) and most queries were hanging.
Is there any reason why?
Note the screenshot shows that 3 deployments were made (2 during the first spike), one at the end of the last spike.
But in fact, the first 2 deployments where not pushed to the bugged environment so shouldn't have an effect on the CPU usage.
The last deployment was made to the environment and seems to have fixed the issue for whatever reason.
Other than today, the CPU usage was super stable before that.
9 Replies
Project ID:
feb68be1-8c18-4a6d-8f6c-072f2338895bfeb68be1-8c18-4a6d-8f6c-072f2338895b
This is very likely a code issue. No incidents have been raised on Railway’s end
Strange because the code has been running flawlessly for multiple days with no changes
Is your traffic routed through cloudflare? It sounds to me like you got a lot of traffic/your code got overwhelmed and you crashed the db
Railway itsself could not have been the cause of this issue. All the platform does is host the code that you wrote
Yep routed through cloudflare indeed. Using the same code on multiple providers with even bigger traffic and no issues encountered before
Issue is happening again..
I am hosting an app not a db on railway. I have multiple replicas so even if one server was to crash I expect for the other replicas to pick up
As you can see on the screenshot, there were no major issues before yesterday
No traffic spike either during the last 2 days
Yeah I get it, obviously. It just seems strange that it happens suddenly the first time now. We didn't encounter any issues on Vercel (serverless so not really comparable) but also render and fly
Interesting, you’re on the pro plan so you can reach out to the team directly through [email protected]
Make sure to include all the details mentioned here as well as your project id and account email