Right now the tcp output plugin can hemorrhage events if the remote side is unreachable or terminates a connection and becomes unreachable. The plugin properly handles reconnecting, however there is a period of 10s at minimum where events will be lost.
The idea would be to re-queue the event until the plugin becomes operable again, however in the event that there are multiple outputs this could cause duplication to some if not all of the other configured output plugins. This should be handled gracefully.
Right now the tcp output plugin can hemorrhage events if the remote side is unreachable or terminates a connection and becomes unreachable. The plugin properly handles reconnecting, however there is a period of
10sat minimum where events will be lost.The idea would be to re-queue the event until the plugin becomes operable again, however in the event that there are multiple outputs this could cause duplication to some if not all of the other configured output plugins. This should be handled gracefully.