We have a single mongod container running using the official docker image (we've tried both mongo:4.4.14 and mongo:5.0.8) that at some point stops accepting new connections and any connection requests just time out.

We start the container via docker-compose with:

  mongo:
    image: mongo:5.0.8
    container_name: mongo_database
    command: mongod
    ports:
      - 27017:27017
    volumes:
      - ./dbs/mongo:/data/db
    logging:
      driver: "none" 

This issue occurs every 12-24 hours, but so far we've noticed that if we comment out the last two lines above

#   logging:       
#     driver: "none" 

then it becomes much more rare. I've attached the stack trace collected after the server has stopped accepting any new connections. Configuring mongod to write its log to a file has the same effect of reducing the rate of failure considerably.

I've attached the stack trace after the server has stopped accepting new connections.

gdb_2022-05-31_09-41-54.txt