health monitor issues.
We are currently load balancing our app servers using the F5 boxes running v10.2.0.We tried using multiple health monitors but they failed / ineffective due to the following reasons. the app servers are running on ATG/Jboss. Please let me know if anyone faced this issue earlier and resolved it.
Http - Creating n number of sessions. The session created for the health monitor check do not terminate.
tcp - works fine under less traffic. But when performing load test or high traffic, the box flips the nodes active/inactive within seconds causing error in the site.
External Health monitor - We are using SNMP and this is causing the CPU to be very high. Currently we just enabled the health monitor for one brand and its running at 79% avg. cpu. (We have five more brands to go)
inband monitor - is not effective as it marks the node active even when the node is disabled.