Open XinleiYang7 opened 5 years ago
If the queue limit is reached, then the server is most likely disconnected or it does not have time to process incoming logs and an error message to send to it meaningless. Messages in any case need to be deleted, because it is impossible to make a queue of infinite, otherwise there is a chance to spend all the RAM. In addition, the server does not need to know whether the queue is full on the client or not, because it still will not be able to somehow influence this situation. In addition, sending a message of overcrowding is also added to the queue, which further aggravates the situation. In the aggregate, I think that the current implementation is the most correct.
sending the errors message to server when the errors reach maximum length capacity instead of dropping the oldest within the interval.