Automatic Scaling not working
Project Id: 70ab406f-186f-40f7-ac61-c1a8dcb4bbb9
I am a user on the Hobby plan, the scaling on my project was working fine, i had an expired credit card and ended up losing the subscription for a few days. My subscription is currently active again but now scaling on my project doesn't work, in the Metrics panel i can see that the Memory usage isn't going over 512MB where previously it would go above it by about 5-6x. My service deployment on this project is now constantly crashing probably due to the limited memory is my guess.
14 Replies
Project ID:
70ab406f-186f-40f7-ac61-c1a8dcb4bbb9have you done a full redeploy of the service?
It's not a good option for me on my use case as this is a meilisearch deploy. Looks more like an issue on the account allowed quota on railway side
there is no way to allocate more resources to your service without having to redeploy
you're talking about automatic scaling? railway doesn't automatically scale any resources
I guess it's not really automatic scaling, it's just allocation of resources. I am on the hobby plan which means that "Your plan includes up to 8 GB of RAM, 8 vCPU, and 100 GB of Shared Disk.". But i can clearly see a cap of 512 MB of RAM on my project which is the cap of the free tier/trial tier.
This doesn't make much sense to me unless i am missing something... If a trial user creates a project where he's capped to 512MB and then he needs to upgrade to get more RAM memory he would need to redeploy the entire thing? Don't think this is what happened to me
you where bumped down to the trial resources when your payment declined, to get your deployment back up to the hobby resources you will have to redeploy
I see this same question all the time
"I've upgraded to hobby but can't use more than 512mb of ram"
the answer being, simply redeploy. railway can not allocate extra resources to a running docker image
Got it, ok thanks
same with a hobby to pro upgrade, you need to redoploy your service to get your service to use the pro resources
Looks like an opportunity for improving documentation here if this really happens often because currently this info is not clear anywhere
Thanks for the answer though
you can't add extra cpu or memory to a running deployment the same way you can't swap out your cpu or add extra memory to a running computer
and yes 100% agree the documentation for this could be so much better, but if i recall correctly they want to improve the upgrade flow making it self explanatory so there's no room for confusion in the first place
Yeah this makes sense to me, it's just pretty confusing to me since i didn't have to redeploy it when i first upgraded and it all worked. Only option for this to have worked out is if i was on the hobby plan since the start which i am not 100% sure but i don't think was the case here
but anyway, it's fine, thank you!
a deployment from a code change would have also put you on the hobby resources