Scroll to navigation

tcpconnlat(8) System Manager's Manual tcpconnlat(8)


tcpconnlat - Trace TCP active connection latency. Uses Linux eBPF/bcc.


tcpconnlat [-h] [-t] [-p PID] [-L] [-4 | -6] [-v] [min_ms]


This tool traces active TCP connections (eg, via a connect() syscall), and shows the latency (time) for the connection as measured locally: the time from SYN sent to the response packet. This is a useful performance metric that typically spans kernel TCP/IP processing and the network round trip time (not application runtime).

All connection attempts are traced, even if they ultimately fail (RST packet in response).

This tool works by use of kernel dynamic tracing of TCP/IP functions, and will need updating to match any changes to these functions. This tool should be updated in the future to use static tracepoints, once they are available.

Since this uses BPF, only the root user can use this tool.


CONFIG_BPF and bcc.


Print usage message.
Include a timestamp column.
Trace this process ID only (filtered in-kernel).
Include a LPORT column.
Trace IPv4 family only.
Trace IPv6 family only.
Print the resulting BPF program, for debugging purposes.
Minimum duration to trace, in milliseconds.


# tcpconnlat
# tcpconnlat -t
# tcpconnlat -p 181
# tcpconnlat -L
# tcpconnlat -4
# tcpconnlat -6
# tcpconnlat 10
# tcpconnlat -v


Time of the response packet, in seconds.
Process ID that initiated the connection.
Process name that initiated the connection.
IP address family (4 or 6).
Source IP address.
Destination IP address.
Source port
Destination port
The time from when a TCP connect was issued (measured in-kernel) to when a response packet was received for this connection (can be SYN,ACK, or RST, etc). This time spans kernel to kernel latency, involving kernel TCP/IP processing and the network round trip in between. This typically does not include time spent by the application processing the new connection.


This traces the kernel tcp_v[46]_connect functions and prints output for each event. As the rate of this is generally expected to be low (< 1000/s), the overhead is also expected to be negligible. If you have an application that is calling a high rate of connects()s, such as a proxy server, then test and understand this overhead before use.


This is from bcc.

Also look in the bcc distribution for a companion _examples.txt file containing example usage, output, and commentary for this tool.




Unstable - in development.


Brendan Gregg


tcpconnect(8), tcpaccept(8), funccount(8), tcpdump(8)

2016-02-19 USER COMMANDS