How does this estimation work?
I'm a bit confused by this panel, and why it is quoting $105 estimated bill, yet the service (which is our main server) below, it is saying an estimated $552. How do these figures work? Is there something wrong with the display?
32 Replies
Project ID:
N/AN/A
mind showing the goalsumo breakdown of costs?
I'm suspecting egress discount
I'm not sure I understand what egress means if I'm being honest, I am in over my head on the server administration stuff.
and the stuff over here?
the egress is the traffic your app sends
so if it sends a 3gb file to a user then your egress there would be 3gb
yea seems like it's just your egress discount in effect, probably, idk
So should I be worried about a $500 bill showing up?
yea, if it is your egress discount causing the difference then I'd be worried, the egress discount is not gonna last for much longer, it's only a temporary thing railway implemented
if you guys aren't expecting that much egress then I'd suggest finding the cause of it, things like using a database over public url or using an external file storage solution can contribute to this
if you guys are expecting that much egress then you'd need to make sure the new pricing will fit in your budget, egress was like 60 times cheaper due to a bug before, which is why it's discounted right now ( to give users time to prepare )
it's also possible that it's just a visual bug and I'm completely off the mark here, probably best to wait for someone smarter than I am to pitch in
I think this might be caused by opentel logging, I had to implement opentel logging over HTTP due to not having any native logging soutions... I don't understand how we are sending out 500+ gb of data a month right now, we only serve maybe 600 users a month at most.
And we don't send files via this app, we serve them via S3.
Is your database on railway v2?
the legacy databases are known to cause a egress usage
( you can tell whether it's legacy or v2 by checking whether or not it has a volume attached to it, if it has a volume attached then it's v2 )
No it is on v1. Would it cause something like that? This is all so strange, our production database is not even a gig large.
possibly, you should definitally migrate to v2 asap
Railway Docs
Migrating from Legacy Plugins to Database Services | Railway Docs
Documentation for Railway
Thanks a lot for your help @Fragly .
Railway Team, if you can follow back up on this, something is really not right about our traffic trends from the past few days, and I'm not quite sure what's going on....
We had basically no network usage for the longest time (as far as I know, the graph data only goes back 7 days, but I don't remember ever looking at network usage and thinking "that's high"), and after a restart today, we went back down to having no network usage, along with an increasing amount of network flowing out from our app every day, with no increase of users or feature usage to our website according to our product analytics. And then as soon as I restarted the app (I restarted after I saw really high RAM usage), it drops back down like nothing ever happened. I removed my HTTP logger just in case that is what was causing issues.
I can see bad code causing the high RAM usage, but I don't have any idea where our app would have sent 500gb of traffic to over the past month.
I will be upgrading the database tomorrow to v2 as well.
hey deani can you please link your discord account to railway
I think I am linked
link discord here https://railway.app/account
Should I remove and re-add?
yes please
Done. It's not letting me join priority boarding again, but I'll put in a different support ticket for that tomorrow (it's almost 4 am here)
This is the relevant project ID, since I didn't add one to this ticket before: 29a5fb27-3661-4597-af03-ec799c4720cd
the bot isnt letting me flag this thread so ill have to tag @matt directly - can you look into the network usage
If you’re on a legacy DB it’s 99% your DB connection being external
Upgrade, use the internal URL, and report back
In the future we should mention that network, like memory and CPU, is all on the users application
And that v1 plugins and public URLs are the likely culprit, so they should upgrade and use private URLs
CC @matt
Will take care of the DB migration tomorrow and update the ticket. Was this not previously reported in the metrics data before V2 came out? Restarting the app server led to the network going back down to pretty much 0.
We were accidentally undercounting our network by 58x. You should have seen this in an email
We’re waiving it 100% for legacy plugins this month, and 50% in Dec
Lol, that sounds like something I would have been responsible for if I worked for Railway. I hate server work. Still, our network usage went back down to 1/500th of what it was after a server restart, which is still kind of confusing. I will update the ticket after my DB is updated tomorrow and I get some sleep, good night folks 🙂
Restart going down is confusing. Should that pop up more, we will defs have a look
As it stands tho, the calculations are correct, it would just be a cosmetic thing on the metrics tab
Is there a way that we can cap our network usage in the scenario of another anomaly? I don’t want our company to be liable for hundreds/thousands of dollars in hosting bills. This event looks like it’s going to cost us some $$$ if it were to happen in the month of December.
Yes you can set spend limits check the docs
Without any changes to my database version or anything, traffic after the restart has stayed down without any spikes. I will still handle a migration to v2 database, but I don't see how this was related to the database being over used.
