All the nodes suddenly failed to register with the hub (periodically, roughly every 30 minitues)

[kqiang]

We have a monitoring system to monitor how many nodes are registered with the hub. It checks every minute with this command:
curl -s -w 'STATUS%{http_code}' --max-time 30 http://hubip:4444/grid/console# | grep 'host:' | wc -l

After upgraded selenium grid extras to the latest version (2.0.4), we started to see that all the nodes suddenly failed to register with the hub (i.e. the above command returns 0 nodes registered with the hub, it might be due to cannot connect to hub at all), roughly every 30 minutes. And every time it recovered by itself in 1 minute.

I also tested all the grid extras 2.X.X version, they all have the same issue. I was using 2.12.12 before, everything worked well with it.

Did anyone else encounter the same issue?

[kqiang]

The further investigation indicates that we lost connection to hub roughly every 30 minutes. The nodes are still registered with the hub, it's just we lot the connection to the hub. This lasts 1 or 2 minutes and after that we can still detect all the nodes.