cancel
Showing results for 
Search instead for 
Did you mean: 

Custom HTTP monitor query

Gajji
Cirrus
Cirrus

If configured custom HTTP monitor with send string check on the server

 

would that HTTP request will go from both the Active-Standby loadbalancer to the server? any if both loadbalancer send requests they why?

 

if if the URL is dynamic will that HTTP request from the load balancer to the server will not close immediately??

3 REPLIES 3

Hi Gajji,

 

yes, both loadbalancers will send the health checks to the server. They do this, so that both of them individually "know" the health status of the server. If the standby device would not know the health status in case of a failover event, it might send traffic to a server that id down.

 

For the "dynamic" URL I am not quite sure what you mean. Something like "/admin/status.php?appID=02"?

I think it will not differ from health checking a static URI. If you have a proper receive string it should work.

 

KR

Daniel

Thanks, Daniel for the Reply...

 

Actually issue was wait time more for the request from LB to backend...it's not getting closed immediately...

 

The application team said its a dynamic URL... so I just change the path from /monitor/monitor

too simple "/" as per their suggestion and wait time was reduced

It depends on the value you are getting from this content check monitor. If it helps you to reliably determine the server state, you should discuss with the application team why the response takes so long. In the end it should be their interest as well that the BIG-IP can quickly check the state of the server and does not impose additional load on the server by querying dymanic URLs.

 

You could tune your health monitor too.

With "time curl http://backend/status.php?appID=02" you can measure the time the server needs to respond and adjust your monitor.

Or you check whether an adaptive monitor or an EAV external monitor might be of use for your case.