In my last posting, I wrote about a Zuul service becoming unresponsive due to the death of a Log4J thread. Here, I will continue our discussion and try to arrive at a resolution to prevent the reoccurrence of the bug in the future.
Recently, I came across a production issue where a Zuul (Spring Cloud Brixton.SR3) edge service became unresponsive to HTTP requests. Even the health REST endpoint, which provides basic health information, of the affected service failed to respond. The Eureka registry service also advertised the unresponsive service as up. It was a Spring Boot (1.3.4.RELEASE) application with an embedded Jetty (9.3.0.v20150612 ) web container running on Red Hat Linux VM with Oracle JVM (1.8.0_31).