Forum Discussion

Shadow's avatar
Shadow
Icon for Cirrus rankCirrus
Dec 05, 2023
Solved

Health monitor question

We have a user who is seeing this message very frequently in their logs from the F5 statis self IP. 10.192.10.XXX - - [04/Dec/2023:15:56:36 -0600] "GET / " 404 682 All our nodes are reporting as up...
application delivery
  • PhatANhappy's avatar
    PhatANhappy
    Dec 06, 2023

    A base HTTP(s) monitor will send a " GET / "  to the service port of the pool memember, but does NOT have any recieve string.  ANY response from a server   200/301/404 etc.... is a response to the monitor, and hence the node is marked up.
    in creating a new monitor based off the https monitor, you can keep the default settings and change the recv string to    HTTP/1\.(0|1) (2|3)      <---- that will allow only 200 and 300 errors and 400/500 errors will be marked as node down.   This is a great step in the right direction of providing better quality of service.   It will also in the long run will have you learning all the 400/500 error codes and what they mean as you debug why the pool member is down.

     

     

     

PhatANhappy's avatar
PhatANhappy
Icon for MVP rankMVP
Dec 06, 2023

A base HTTP(s) monitor will send a " GET / "  to the service port of the pool memember, but does NOT have any recieve string.  ANY response from a server   200/301/404 etc.... is a response to the monitor, and hence the node is marked up.
in creating a new monitor based off the https monitor, you can keep the default settings and change the recv string to    HTTP/1\.(0|1) (2|3)      <---- that will allow only 200 and 300 errors and 400/500 errors will be marked as node down.   This is a great step in the right direction of providing better quality of service.   It will also in the long run will have you learning all the 400/500 error codes and what they mean as you debug why the pool member is down.