Hi All!
When installing HA without Docker, when > 1 instance is used, then in all services it was possible to specify:
DISABLE_PRESENCE_MONITOR=true and do not use this parameter in one service, then the statuses were read and changed correctly according to the AutoAway settings and there were no performance problems.
When using Docker containers - this no longer works, the statuses do not change if the user was online (it does not matter if he left, closed the application (electron), etc) - the status does not change.
The user always remains Online, while sometimes away is activated (very rarely), but does not change back to Online (expected behavior)
Resetting sessions, as indicated by the link above, does not help, the user becomes Online again. Yes, his application is running, but the computer is locked and no one is working on it.
In the RocketChat logs I see:
But for these users, the status does not change, it is still displayed as Online, in the mongo UserSessions table - Online is also displayed.
I also tried to restore the production mongo db on the test RC, and for the past 3 weeks, 35 users have had the Online status, but no one has access to this test server :smiley: .
In the logs of the Docker container, I periodically observe the following:
Description:
Hi All! When installing HA without Docker, when > 1 instance is used, then in all services it was possible to specify:
DISABLE_PRESENCE_MONITOR=true and do not use this parameter in one service, then the statuses were read and changed correctly according to the AutoAway settings and there were no performance problems.
When using Docker containers - this no longer works, the statuses do not change if the user was online (it does not matter if he left, closed the application (electron), etc) - the status does not change. The user always remains Online, while sometimes away is activated (very rarely), but does not change back to Online (expected behavior)
my docker-compose.yml
.env
db.usersSessions.deleteMany({}) db.users.updateMany({status:"online"},{$set:{status:"offline"}})
Resetting sessions, as indicated by the link above, does not help, the user becomes Online again. Yes, his application is running, but the computer is locked and no one is working on it. In the RocketChat logs I see:
But for these users, the status does not change, it is still displayed as Online, in the mongo UserSessions table - Online is also displayed.
I also tried to restore the production mongo db on the test RC, and for the past 3 weeks, 35 users have had the Online status, but no one has access to this test server :smiley: .
In the logs of the Docker container, I periodically observe the following:
This problem is observed only in Docker, everything works correctly in a stand-alone installation without using containers.
This bug reproduces on 3., 4., 5.* versions and Docker only.
Nginx config from rocket docs with upstreams.
Thanks for any help!
Steps to reproduce:
Expected behavior:
Must be idle time-out and userpresence changed to away
Actual behavior:
Server Setup Information:
Client Setup Information
Additional context
Relevant logs: