2 days ago
Hello I'm new to railway so I might probably ask stupid question, I have a basic pg + expressjs + nextjs app, this is not in prod yet but I have an e2e suite I run a few times, usually it works fine but sometime the db fails for some unknown reason, this is the only thing I can see in the pg logs:
2025-08-01 23:58:38.693 UTC [74] LOG: checkpoint starting: tim
2025-08-01 23:58:53.210 UTC [74] LOG: checkpoint complete: wrote 148 buffers (0.9%); 0 WAL file(s) added, 0 removed, 0 recycled; write=14.448 s, sync=0.030 s, total=14.518 s; sync files=107, longest=0.015 s, average=0.001 s; distance=596 kB, estimate=1396 kB; lsn=0/2374760, redo lsn=0/2374728the metrics are ok, no spike and really low values, any idea what's going on? locally I never had any issue
1 Replies
2 days ago
If I remember correctly, the free plans have serverless mode enabled by default. I am not sure if you can change that.
I am assuming your nextjs service talks to expressjs which talks to postgres. So what serverless mode does is if no network activity is made in the last 10 minutes then the server shuts down.
What happens here is after 10 minutes of inactivity your nextjs and express app shuts down, then after another 10 minutes your postgres server shuts down.
Is this the accurate behavior ?
btw your db log looks fine.
Solution to this problem would be to turn off serverless mode for postgres. If you can't then upgrade to a Hobby tier.
No in theory is not serverless, in my dashboard is disabled, in any case it usually happens during the e2e, so is not after an idle period…