THIS IS A MAJOR INCIDENT. We’re seeing severe latency spikes every 4 minutes, with no traffic changes on our side. This strongly looks like a host-level issue.

We tried scaling out replicas, but it did not help.

This exact repeated spike pattern has happened before (Railway incident).

Could someone from the Railway team please look into this urgently? This looks like host-level or platform-level degradation, and we are losing tons of money right now because of it.

Do you have performance logging in place to isolate exactly where the latency is originating from? Would be helpful for isolating the cause

At the moment, it looks DB-related from our side. The latency is showing up mostly inside ActiveRecord / PostgreSQL queries, including normally fast indexed lookups. We're not seeing any corresponding deployments, config changes, or traffic changes that would explain this.

Also, scaling out app replicas did not improve the issue, which makes us suspect the bottleneck is not in the Rails app layer itself, but somewhere around the database or underlying host/network path.

We're still investigating, but based on what we can see so far, the slowdown appears to be happening around DB access.

thanks for sharing, this is very useful information, regarding the PostgreSQL database - did you deploy the Railway provided one (Right Click -> Database -> PostgreSQL) or is this a custom one / template deployment?

I used the Railway provided Postgres

Thank you, I'm asking some more people about this 🙏

Thaks. I hope this can be resolved soon.

with the additional reports, I'll escalate this to the team, thanks for flagging guys!