N/A

how is your app connecting to those services?

make sure they are using the private network

https://docs.railway.com/reference/private-networking

Isn't that what this line is doing: 'HOST': 'postgres.railway.internal', ?

Which service causes that egress?

Yep, that's it.
If private networking is properly set between your services then you should investigate what is causing this high egress costs, what service uses that much egress?

I've got no idea. I don't really know what egress even is. I just saw it on my bill. I had this problem before and I started using the internal hosts and it solved it. But even using the internal hosts isn't working anymore.

project id please

6d42db87-7a88-4e8b-85fe-1365cf208831

are you connecting to s3?

yeah

uploading to s3 is egress

just uploading?

yes just uploading, egress is outbound traffic

we dont charge for ingress / inbound traffic

the users can upload their images and videos but they don't do it that often

does your app modify the images or videos? or just pass them onto the s3 bucket?

if you just pass the files stright to s3, then you would want to instead be returning a signed url to the users to upload the files to, that will eliminate the egress of you sending the files to s3

It resizes their images but I've just checked and there are only 6 images in the bucket

alright then maybe you are still connecting to postgres and redis publicly

remove both of their public tcp proxies and redeploy both of them, then check if your app still works

do you have a link on how to do that?

alright, that seems to have worked

I'll just keep an eye on it over the next 24 hours to see if it takes off again

thanks for the help

if you still have egress then it's not egress from the databases, and if so, there won't be much we can help with tbh

you have not removed the tcp proxy from postgres

Alright, yeah, I've just noticed that

it's weird that it keeps on using the public connection

the costs have definitely come down

pretty sure it was the Redis connection that was doing it as that it is updated multiple times a second

hey, how is it going with the egress?

Since the changes were made, it's gone up by 10gb

then that would be genuine traffic, it's not from the database given there's no TCP proxy anymore

Where could that be coming from though?

I'm sorry but I have no way to tell you with any level of certainty, we do not track our users internet usage in such a way that would let me tell you anything other than how much you're using (though of course you can also see that yourself)

I had another look at this today and it's added another 20GB of egress. What else other than database connections can cause this?

I am running Celery and Redis and Celery is making API calls to get additional information. Can this be a cause as well?

any outbound traffic counts as egress

Guess that explains it then

!s