Slow performance for Pro Plan
I don't know if this is expected behaviour but I'm experiencing slow performance for each build, deploy steps and runtime.
Using github actions and deploying my app to own decent dedicated server 12vCPU, 32G RAM my build takes like 30s. Using railway it takes like 7 minutes. I'm executing additional tasks beside
yarn start like dropping tables, creating new tables, seeding database and so on so maybe database is the problem?
Runtime is also really bad. Simple GET request which on my dedicated server is taking 60ms and on AWS t3a.small is taking like 120ms on railway take 900ms.
I'm a little bit confused since with Pro Plan I should get 32 GB RAM / 32 vCPU per Service but I don't feel this.12 Replies
Project ID:
a25b5b70-0a40-4922-afef-8905f1b71ea6Project ID:
a25b5b70-0a40-4922-afef-8905f1b71ea6It depends where you are doing the GET from on your server and from where railway is doing it
By region difference this will have a big impact
I'm making requests from Poland. My dedicated server and AWS region is located in Germany (eu-east). Railway app is deployed to eu-west region. I don't think region is problem here since eu-west is a little bit further and build and deploy steps should not be affected by my location
i was only referring to the GET
@kierzniak What db do you use and where is it hosted?
MySQL . I have just find out something. When I joined beta program in all my services there was a select to change region. In MySQL there was not such a select. I have removed MySQL service and add it again and I see that I can change region of MySQL database. I think that might be a cause
Alright, you can test that. I'm using a postgres myself in a pro plan hosted on Railway and I experience slow execution on my queries, not included any latency.
Probably problem was the MySQL database service which I could not create in my eu-west region. Automaticly services are created in default region and then we can migrate them to our region. I had a problem with migrating MySQL volume so I give up with railway database and attach my database server. After switching database deployment and runtme has decent performance
How many SQL queries are executed in the GET request? Latency with DB connections can have significant impacts depending on the app.
Take WordPress for example. Add 100ms latency to the DB layer and 500ms FTTB will become 10s pretty quickly. (WordPress executes 100-300 SQL queries per request.)
Sure, this time it was database but in not the same region. Since I'm using dedicated database in the same region performance is decent
btw. how I can mark this post as solved?
right click > apps > mark solution