New function: TCP connection monitoring

[jundizhou]

Describe the solution you'd like
Monitor whether the TCP connection is successful.
Monitor the cause of TCP connection failure.
Monitoring TCP connection time.
Describe alternatives you've considered
I will provide two kinds of events at the bottom:
First, TCP_ Connect. The occurrence of this event indicates that a TCP connection has occurred. It can be used together with syscall-connect to determine the starting point of analysis. At the same time, the event will also give some reasons for the failure of the company
Second, TCP_ finish_ Connect. This event only appears when the connection is successfully established. You can judge whether the connection is successful or not through this event
Additional context
If anyone is interested, you can supplement the code in the collector and grafana-dashboard

[jundizhou]

This is to solve a scene in the capsule

[jundizhou]

I tried to code for it. I found that it is impossible to judge whether the connection is successful only from the client, because the client defines that it can send data by completing the second handshake. The connection success can be judged only from the server, but the connection time cannot be calculated. Therefore, the feasibility of this function is in question

[jundizhou]

If the ACK transmission of the third handshake fails, resulting in the server's delay in receiving the ACK, the resource will be released. At this time, the client thinks that it has been connected, it will send data to the server. Because the server does not receive the third handshake, it will respond to the client with RST packet.