Hey @roofboard , are you using Baserow 1.10.0?
The first thing you could try is setting the
BASEROW_AMOUNT_OF_GUNICORN_WORKERS to a larger value, it defaults to 3.
Secondly are you able to provide any resource usage graphs of Baserow during these peaks showing cpu/memory usage? And the same for the postgres provided by cloudron and used by Baserow (if available).
Could you give more details on how exactly you are interacting with Baserow using n8n? Are the majority of these requests just row reads or are they writes also?
I believe the current n8n Baserow integration is non-optimal also and could be causing performance issues. Specifically I believe every time n8n fetches rows from Baserow it runs a separate request to get all the field metadata. This isn’t needed anymore due to the new
user_field_names flag, however this would require a change to the n8n Baserow integration to fix.