How does the inband monitor work?

Hello.

You can capture the traffic of the HTTP monitor using tcpdump.

# tcpdump -vs0 -nnei 0.0 host <backend_server_ip> -c 10

Or even better, using this (to check with hexadecimal info)

# tcpdump -Xs0 -nnei 0.0 host <backend_server_ip> -c 10

KR,

Dario.

The inband monitor is a completely different kind of monitor from the other monitors offered by BIG-IP in that it doesn't send any content to a pool member itself (it's not whats termed an 'active monitor').

Instead, it looks at the success of connection setup for load balanced traffic - if a SYN is sent to a server and the server doesn't respond to with a SYN/ACK, that's an error and if the number of errors in a rolling time interval breaks a threshold, the server is marked down. The inband monitor will release the pool member from it's down status after a period of time (300 seconds by default), and begin tabulating errors again.

I have a pool that contains a single server and when I assign the inband monitor to the pool it will take the server down for short periods. However when I create an https monitor the pool member will stay up and not report any errors. I thought that if the inband monitor failed all other monitors would not function properly. Would it be possible to explain why this might be? Thanks