Node.js endpoint execution stops without reason.
Hello everybody, I have this function that works perfectly fine on my local environment, but for some reason it doesn't here on railway: I have this webhook that receives data and pass it to a class that could take up to 20minutes to process data, like this:
Webhook.js -> Utility.js
Problem is, the execution terminates inside Utility.js (without any errors) without passing back the control to Webhook.js, I suppose it's a container/ resource problem (I'm on hobby with 50 dollars of credit), it really preventing me to use my project. Any hints? Thanks
42 Replies
Project ID:
6b0f0506-09fb-47d1-9a76-e917fa00c0046b0f0506-09fb-47d1-9a76-e917fa00c004
do you have a container died event in your logs?
appeared only when I restarted
like container restarted and container died
then if you didn't get one while your app stopped the app didn't exit
that's you initializing it
mm k, no I didn't get it
got stuck on the last log of Utility.js class
are you aware of the resource limits of the Hobby plan?
8gb of RAM + 8v CPU
+ extra resource
where can I see if the request exceded the resource limit?
I have no problem paying more, but I need to know
you can look at the metrics, but your application may have had a resource spike that happened faster than we could have caught
what's the max memory that it uses locally?
I'm running it on 32gb macbook pro
I can do a benchmark if you want
yes please
k
you can always upgrade to pro and If that doesn't solve your problem I'd be happy to refund
sure, but I want to give you the benchmark first
may help others if it's a bug
can I upgrade to pro using credits?
nope it would have to be a card, but we can transfer credits over after the fact for you
mmm, cause right now I have just a pre-paid
you would need to use a card to purchase pro
this is the benchmark as stated here: https://nodejs.org/en/learn/getting-started/profiling
Node.js — Profiling Node.js Applications
Node.js® is a JavaScript runtime built on Chrome's V8 JavaScript engine.
By the way I've DMed you, we can talk better there I think, I'll upgrade, no problem
sorry but we don't offer support over DMs
I understand, sorry 😉
can you give me a rundown of that file? what was the max memory your app uses?
this is the process consumption
zsh is a shell?
yep, I have an instance inside visual studio code
what does 47426 mean in this context?
pid
well isn't this idle usage?
nope, I sent the request on my local instace of the express app
but at the same time, there's no way an idle node app is only using 2.8mb, are you sure that's not the memory usage of zsh itself
Ok, I'm going to try a different method
The script uses approximately 87.61 MB
I used this: const used = process.memoryUsage().heapUsed / 1024 / 1024;
At this point it's should be helpfull to check the logs inside the container
on your side
we ship all the deployment logs your app prints out to deploy logs tab
ok, so any suggestions? Just upgrade?
if you think this is a memory issue, yep!
if you don't think it's memory, I would recommend adding more verbose logging to your app
There are already
I'll try again, thanks!
sounds good
Hi! I re-deployed the env variables a couple of times and it worked, thanks!
awsome