ERR_EMPTY_RESPONSE after 120 seconds after migrating to Openshift 4.10

We are facing issue after migrating our application from Openshift 3.11 to 4.10

App is working just fine on both Openshift versions but when I tried to process some amount data (Databa开发者_JAVA技巧se query) through our app running on Openshift 4.10 + docker, after exactly 120 seconds i get ERR_EMPTY_RESPONSE (it works just fine on old Openshift)

I thought it might be issue with DB but i put 180 seconds sleep to random button within app and after 120 seconds i get same error.

We are using Gunicorn so i have tried to add GUNICORN_TIMEOUT to env vars but this didnt help. I also added timeout argument to gunicorn run command in entrypoint but this also did not help.

We have haproxy.router.openshift.io/timeout set to 3600s in annotations.

Gunicorn Timeout is set to 3600 also.

Any ideas what might cause the issue? We have nowhere in our app code 120 seconds timeout so it might be somewhere else and the fact that it works perfectly fine on Openshift 3.11 with exectly same config makes it even weirder.