Forum Discussion
NimbostratusF5 is sending very frequent health checks to application servers
F5 is sending very frequent health checks to application servers . In last 7 hours LB has sent 20000 health checks to one of the backend server . Can someone suggest is there any issue with the monitor configuration ?
Below is the monitor configuration .
monitor http http-test { adaptive disabled defaults-from http destination : interval 5 ip-dscp 0 recv 200 recv-disable none send "GET /fake-test/services/CRMASYNC/zz/zz_ep?WSDL HTTP/1.1\r\nHost: fake.domain.com\r\nConnection: Close\r\n" time-until-up 0 timeout 16
Stanislas_Piro2Cumulonimbus
Your health check is defined with an 5 seconds interval.
So each cluster member will request each server every 5 seconds (which means 20 times per minutes)
7 hours * 60 minutes * 20 times = 8400So in a 2 members cluster, each server will receive 16800 requests in 7 hours...
If you want to reduce this number, increase interval...
if you increase interval, the failure detection will be longer.
- Daniel_Wolf
MVP
If you are just bothered by the ~20.000 log entries, for most web servers you can exclude certain IP addresses or certain User Agents from being logged. Works for sure with Apache and Nginx.
Just exclude the Self IP Addresses from being logged.
If the additional load caused by these requests is an issue - increase the interval.
