This happens randomly after seconds, hours or even days. Restarting odoo, nginx and postgres services doesn't fix it.

My Odoo log says:

2022-03-01 17:47:58,573 3273 INFO odoo longpolling: 192.168.x.y - - [2022-03-01 17:47:58] "POST /longpolling/poll HTTP/1.0" 500 140 71.253036

Ningx log indicates:

2022/03/01 12:48:06 [error] 2836#2836: *106422 connect() failed (111: Connection refused) while connecting to upstream, client: 192.168.x.y, server: ..., request: "POST /longpolling/poll HTTP/1.1", upstream: "http://127.0.0.1:8072/longpolling/poll", host: "...", referrer: ".../web"

And PostgreSQL log indicates:

2022-03-01 12:53:13.212 -05 [3442] odoo@odoo ERROR:  40001: could not serialize access due to concurrent update
2022-03-01 12:53:13.212 -05 [3442] odoo@odoo LOCATION:  ExecUpdate, nodeModifyTable.c:1442
2022-03-01 12:53:13.212 -05 [3442] odoo@odoo STATEMENT:  UPDATE "bus_presence" SET "last_poll"='2022-03-01 17:53:13',"last_presence"='2022-03-01 17:52:28.073075' WHERE id IN (50)

Odoo is running on Ubuntu 20.04 with 16 GB of RAM and 4 vCPU.

I've configured 5 workers, but I've tried with 0 and 1 and the result is the same - although with 1 worker odoo behaves more stable.

I've also tried changing multi_accept values on nginx with the same results.

Any idea?