Weblogic threadpool queue length
In an incident where some backend systems is down, and we checked weblogic sever health. Clicking on server => monitoring => thread, and we see a table with a column named 'queue length' and the number is high like more than 2000.
According to Weblogic doc, this 'queue length' means pending requests including user and system requests. however, there are plenty of idle threads at the same time. and if we clicking on the 'work load' tab, all pending requests number is 0. and the weblogic servers seems work fine otherwise.
So my question is: does the 'queue length' really mean pending requests? why we have plenty of idle threads while we still seeing high number of pending requests? why the pending requests number does not seem to adding up by looking into the 'work load' tab?