I'm using Faust in combination with a Django backend. Faust is integrated as a Django app but runs in a separate container, similar to how Celery workers operate — except in this case, it's Faust with Kafka.
Inside my Faust agents, I'm performing asynchronous operations using Django's ORM. After the system runs for a while, I begin encountering connection closed errors during ORM operations. This seems related to how database connections are managed in long-lived async processes.
any idea how to handle this?
logs:
920 | psycopg2.InterfaceError: connection already closed |
-- | -- | --
| | 2025-09-12 09:56:15.920 | django.db.utils.InterfaceError: connection already closed |
| | 2025-09-12 09:56:15.920 | psycopg2.InterfaceError: connection already closed |
|
| Fields | | container_name | processor_agent.1.dic96f1ikb3zgicm8nl9l7ih7 | | filename | /var/log/docker/9cd7bc50abb53e57aabebe6f7f7b71df007ae4be86155ac89313e551c0fda5e3/json.log | | host | XXXXXXX | | source | stderr | | swarm_service | processor_agent | | swarm_stack | alertify
Fields
| container_name | processor_agent.1.dic96f1ikb3zgicm8nl9l7ih7
| filename | /var/log/docker/9cd7bc50abb53e57aabebe6f7f7b71df007ae4be86155ac89313e551c0fda5e3/json.log
| host | XXXXXXXXX
| source | stderr
| swarm_service | processor_agent
| swarm_stack | xxxx
| | 2025-09-12 09:56:15.920 | django.db.utils.OperationalError: server closed the connection unexpectedly |
| | 2025-09-12 09:56:15.920 | psycopg2.OperationalError: server closed the connection unexpectedly |
| | 2025-09-12 09:40:40.671 | elastic_transport.ConnectionError: Connection error caused by: ProtocolError(('Connection aborted.', RemoteDisconnected('Remote end closed connection without response'))) |
| | 2025-09-12 09:40:40.671 | urllib3.exceptions.ProtocolError: ('Connection aborted.', RemoteDisconnected('Remote end closed connection without response')) |
| | 2025-09-12 09:40:40.671 | raise RemoteDisconnected("Remote end closed connection without" |
| | 2025-09-12 09:40:40.671 | http.client.RemoteDisconnected: Remote end closed connection without response |
| | 2025-09-12 09:40:40.671 | raise RemoteDisconnected("Remote end closed connection without" |
| | 2025-09-12 09:00:02.668 | elastic_transport.ConnectionError: Connection error caused by: ProtocolError(('Connection aborted.', RemoteDisconnected('Remote end closed connection without response'))) |
| | 2025-09-12 09:00:02.668 | urllib3.exceptions.ProtocolError: ('Connection aborted.', RemoteDisconnected('Remote end closed connection without response')) |
| | 2025-09-12 09:00:02.668 | raise RemoteDisconnected("Remote end closed connection without" |
| | 2025-09-12 09:00:02.668 | http.client.RemoteDisconnected: Remote end closed connection without response |
| | 2025-09-12 09:00:02.668 | raise RemoteDisconnected("Remote end closed connection without" |
| | 2025-09-12 09:00:02.428 | django.db.utils.OperationalError: server closed the connection unexpectedly |
| | 2025-09-12 09:00:02.428 | psycopg2.OperationalError: server closed the connection unexpectedly |
| | 2025-09-12 09:00:02.428 | [2025-09-12 10:00:02,374] [7] [ERROR] 🔥 Error processing alert 'XXXXXXXXX': server closed the connection unexpectedly |
| | 2025-09-12 09:00:02.346 | psycopg2.InterfaceError: connection already closed |
| | 2025-09-12 09:00:02.346 | django.db.utils.InterfaceError: connection already closed
I'm using Faust in combination with a Django backend. Faust is integrated as a Django app but runs in a separate container, similar to how Celery workers operate — except in this case, it's Faust with Kafka.
Inside my Faust agents, I'm performing asynchronous operations using Django's ORM. After the system runs for a while, I begin encountering connection closed errors during ORM operations. This seems related to how database connections are managed in long-lived async processes.
any idea how to handle this?
logs: