Horizontal scaling is here! This is one that we’ve been 

, so we’re ecstatic to get this feature into your hands. 

You can now set number of replicas in your service settings

Vertical scaling only goes so far. At a certain point your application runs out of CPU or RAM and then, as they say, what do?

Previously on Railway we offered a workaround 

 but our goal was always to have horizontal scaling out of the box. 

Horizontal scaling on Railway has a number of benefits:

Eliminates a single point of failure for your application, thereby improving availability

Reduces cost over time as vertical scaling gets expense faster than horizontal scaling

Enables scaling your application to infinity and beyond

If you’re wondering about load balancing, in this rollout we’re doing round robin load balancing. Since Railway VMs under the hood are similar, the risk of federating a request to a server with less vertical capacity is moot.

 panel in your service where you should now see a 

We’re in the process of migrating our logging service over to 

 to give you the speedy logs that you so desire.

Performance so far is great but we’re slow-rolling this migration to make sure that there are no anomalies. 

We cleaned up some lingering CSS issues leftover from the 

We added better explanation text when deleting a successful deployment

We fixed an issue with dates related to the 

We fixed a bug with scroll locking in logs

https://res.cloudinary.com/railway/image/upload/v1684534215/replicas_llmtci.png

Changelog #0134