konecty/meteor-user-presence

UserPresenceMonitor.onSetUserStatus triggering much more frequently than expected

Opened this issue · 1 comments

We have console.logs inside this callback that are trigger log velocity warnings sometimes over 25k lines of log messages within a minute. Our server has 12 cores and we are using meteorhacks:cluster. As far as I can tell we have correctly set up meteor-multiple-instances-status as directed. It's working as expected but I'm concerned about all the excess activity in the logs. Most of the log messages show that the user's status hasn't actually changed.

Here's a small sample from our logs

 Updating user status for User: ggXFMHvthnTTGQj6v
 Previous status: online
 new status: online
 status connection: online
 User is online so clearing users timeout warning
 Updating user status for User: ggXFMHvthnTTGQj6v
 Previous status: online
 new status: online
 status connection: online
 User is online so clearing users timeout warning
 Updating user status for User: ggXFMHvthnTTGQj6v
 Previous status: online
 new status: online
 status connection: online
 User is online so clearing users timeout warning
 Updating user status for User: ggXFMHvthnTTGQj6v
 Previous status: online
 new status: online
 status connection: online
 User is online so clearing users timeout warning
 Updating user status for User: ggXFMHvthnTTGQj6v
 Previous status: online
 new status: online
 status connection: online
 User is online so clearing users timeout warning
 Updating user status for User: ggXFMHvthnTTGQj6v
 Previous status: online
 new status: online

As you can see here all of these triggered for the same user and the status wasn't even changing. Sometimes we have thousands of lines like this all for the same user in our logs.

I am having the same problem!