Random app crash, no logs or other indicator
When trying my healtcheck I get random crashes, no recent deploy was made, I don't see any logs on my systems nor in Railway Logs https://api.cretia.app/.well-known/apollo/server-health
22 Replies
Project ID:
97615b19-20a7-4b88-887d-d4df0fa9a3c1
97615b19-20a7-4b88-887d-d4df0fa9a3c1
Also, railway healtchecks seem to be passing
a 503 is not always indicative of a crash, it usually means your app is listening on the wrong host and or port, please read this
https://docs.railway.app/troubleshoot/fixing-common-errors
It has been working just fine the past couple of months
First time that I have this issue
🤔
Will check, thank you
If that would be the case, wouldn't I get a consistent
503
? instead of being it random?
My configuration looks fine
🤔what kind of app is this?
Node + Docker
I'm fine tunning the ARG in the Dockerfile
Hopefully that helps
well its railway so of course its docker
the ARG for what?
PORT
you dont need a PORT ARG
send the dockerfile please
Looks like it's working now anyway
Thanks for your kind help!
like i said, you dont need ARG PORT or even ENV PORT
why are you globally installing npm with npm?
im sorry but theres a lot wrong with that dockerfile, and in the intrest of stopping someone from coming along and using it, im going to have to ask you to remove it
@cesargdm
If you install npm with npm you update it, for security reasons i.e.
my statement stands, sorry
But since I’m not using it for installing dependencies, i would not use that, you are right there
i can remove it if you want?
Why wouldn’t I need the ARG or ENV for PORT?
🧐
you dont need a PORT during your build
the PORT variable is used when the app starts, not during build
I see 🤔
And what about the EXPOSE
Wouldn’t it need to know which port to expose?
you listen on the
PORT
generated by railway, thats the only way railway will know how to proxy requests to your app
i take it you maybe just skimmed that docs page?Probably, let me have a second round to the docs