a3b76fbf-72bb-4131-9e22-b084af93e94a

you should probably use the sigterm event instead of beforeExit

process.on('SIGTERM', () => { console.log("exiting...") });

also Railway force kills your container after a while if it takes way too long to exit, you can set the RAILWAY_DEPLOYMENT_DRAINING_SECONDS to how long you want to prevent that from happening

cool ! thanks

SIGTERM is never logged

And I don't see any changes in my database (should exist if my function "saveUsersProgress" launches after SIGTERM

do you run your app with npm run <..>?

it seems that if you run it with npm, npm will receive the sigterm instead of node because node process is a children of the npm process
info:

yes it stats with npm run

try to start it with node directly then

what should i change then ?

add a build command and then node index.js

(its a ts project)

go to your service settings, scroll down and you should find a custom start command
place your start script command there

service*

Hey @ThallesComH sorry for the delay

it didn't work

i tried looking for those logs, but nothing

@ThallesComH Hey sorry for the delay

it didn't work

no logs when redeploying

none of those

hey, thats really weird
i'll try to write a code to see whats happening, report back here soon

ty

what did you end up setting your start command to in the railway service settings?

likely just an incorrect start command, since thalles is right, this is caused by you running your app through npm as npm does not pass down signals

Just checked back, It seems like I didn't save the changes, did it now and it works

ty so much guys

awsome!