3 months ago
My backend API has been experiencing high network latency, nearly 10x more than usual, when interacting with my Postgres and Redis instances on Railway. I'm using their internal connection strings.
The issue started around 25 minutes ago. I have a hunch the issue comes from Railway since it's affecting my Postgres connection and my Redis connection.
First image is a trace from earlier today. Second image is from just a few minutes ago.
204594ed-2cbb-4f9e-924a-88cf5d199e27
5 Replies
Looks like the latency went back down within the last hour. No correlation with anything that was done on my end. I'd still like some insight on this if possible.
3 months ago
Your service burst to GCP during the time you reported high private network latency, and was put back on Metal when you reported the latency went back down.
I have staged a change to pin that service to Metal. Please deploy the change at your earliest convenience.
It might also be helpful to add a health check. Nothing to do with this, just good practice.
Shit I saw that change and discarded it without checking here first. Sorry about that. A health check is definitely on my todo list though.
3 months ago
I've re-added the change.
3 months ago
I see you have deployed the change, so I'll mark this as solved!
Status changed to Solved brody • 3 months ago