Passivation errors on F5 health check
SummaryPega is running on a two-node cluster with an F5 load balancer in front. F5 performs a health check for each node by addressing the URL: "http://servername:8080" (the Tomcat Manager home page) and expecting an HTTP 200 response. A large number of unauthenticated browser requestors (>700) created by the F5 health check are not passivated for several hours. Only when the health check is set to a simple ping does the issue disappear.
Error MessagesNot Applicable
Steps to ReproduceConfigure an F5 health check to test if a node is up by addressing the URLs given in the Description section.
Root CauseA defect or configuration issue in the operating environment
ResolutionHere’s the explanation for the reported behavior:
The right approach is to use a service (for example, Service-REST) on the Pega side to handle the ping requests when defining health monitor on the F5 side. With Pega 7.1.9 and later releases, Pega comes with a built-in service called monitor.pingService. For Pega 7.1.8 and earlier releases, create your own Service-REST with an implementation similar to monitor.pingService.
Published April 11, 2017 - Updated May 2, 2017