a month ago
Hello!
I have an n8n service that works with a Postgres database deployed on Railway.
From time to time, Postgres database seems to degrade or completely become unavailable.
I've noticed that this causes my workflows to hang indefinitely, and the entire n8n service degrades completely.
{"code":503,"message":"database is not ready!"} appears.
The capacity and size of the Postgres database are normal. Logs and screenshots are attached.
All settings are exactly as they were in the original template I took from Railway.
Please investigate this issue: what could be causing it? Is it possible that Postgres starts to degrade over time, or that some settings are missing?
Because everything is currently set up exactly as it is in the template, which implies it should be running flawlessly.
Postgres logs:
I also noticed that through the Railway UI, the database connection process is endless.
Meaning, it can't connect to show me the main tables.
3 Replies
a month ago
Hey there! We've found the following might help you get unblocked faster:
If you find the answer from one of these, please let us know by solving the thread!
a month ago
These are error logs from n8n.
The tickets you suggested above don't seem to solve my problem or answer the question.
This is because absolutely nothing has changed in the variable settings or links between databases, and overall the connection works correctly intermittently.
That is, it opens and loads everything, but then suddenly stops working after some time. This is a floating error.
I only use PostgreSQL to store n8n's service data.
I don't connect to this database in any way from within n8n itself. It functions exactly as it was originally created using the template.
Only n8n manages it.
So, errors likely related to the program itself are minimal here.
Within n8n, runs occur quite rarely.
Meaning, there are no cyclical algorithms here or anything like that which would heavily burden my database.
24 days ago
Hi! did you ever find this out? I'm having the exact same issue and scaling up workers def doesn't fix it