Hey!
I’m fiddling with a pet project and have it up on a following setup
app: 2 machines of shared-cpu-1x@512MB running rails
db: 2 machines of shared-cpu-1x@256MB running pg-14 on a (twice?) legacy non-flex image.
everything is in the ams region
The app machines are fine with the average load of 0.12 and a 0.5% CPU utilization. Most of the time, the app is pretty much idle, with literally a couple of requests per hour
The db machines, though, are in a state that I can’t understand. Both machines are constantly at almost their baseline of 6.25% CPU utilization. And as soon as any real requests from the app come, everything is getting hammered with throttling. Sometimes it even throttles when the app remains to keep its usual idle pace. The built-in metrics show ~20 queries per second, but ~18 of them are attributed to “postgres” (light blue in the screenshot), and as you can see, there are constant replication lag spikes. Graphana shows that CPU load as well.
I’ve tried replacing machines with new ones, but nothing has changed
I understand that it’s an entry-level machine and such, but to me, it looks like I’m just missing some source of load
So the question is: what’s going on, what are these “postgres” queries, and what is it constantly replicating?
