the current server is the server of the 2-core 4GB
swoole construction. There are 600clients (will continue to increase)
the client sends a heartbeat to the server every minute, so it theoretically receives 600heartbeats at a time and throws them to task for logical processing.
the time taken by one of my processes is probably 200ms
document:
SWOOLE_CPU_NUMSWOOLE_CPU_NUM * 1000
1100ms1000QPS100
40M1004G
taskonTasktaskworkerworker
the biggest thing I can set is task_worker_num= > 4000?
1. Under what circumstances does the delivery capacity exceed the processing capacity? does it mean that the first 600 heartbeats have not been processed before the second heartbeat?
what does it mean that 2.task_max_request and max_request set the maximum number of tasks for a process? Will a new process continue to run after it automatically exits after exceeding the value? Whether all previous connections will be disconnected.