What’s the source of 502s? not enough unicorn workers? We just moved servers and users have been reporting occasional 502s after the move. We have a 6C/16GB server, and set the unicorn_workers to 16 (before we moved server it was at 12), db_shared_buffer to 4GB, db_work_mem to 60MB.
Would having too many unicorn workers cause 502s? (speculating since we had 4 fewer before the server migration)
I am trying to see if its the configuration that we need to tweak or something else that we need to address.
Thanks.