Identify long time calls on apache balancer
Hi,
We have an Apache 2.2 that refers calls to the application which is balanced.
So there a call from apache to its virtual host balancer.
But Most of the calls have a long time.
The request does not fail, it just takes long times to execute. The balancer sends the request to the application (Java) and it executes finely.
However, I cannot see what is causing this long time response for this call.
Any ideas what to instrument or look for in this case?
Or any Apache configuration should we look at?
Best Answer
We had a case where The Apache to Tomcat Connector used the default setting of 5 or 10 threads. This was on one of our early installations of our Community Portal running Confluence. This became the bottleneck as the Connector could only forward 10 requests at the same time. You always have to make sure that the thread pools on all involved components - Web Server + Connector + App Server - are aligned. Its worth checking the settings
Best Answer
Renato,
Are Apache and the Java application on the same system? (I guess not.)
The delay could be in the communication between the two processes. I'd first take a look at the infrastructure monitoring dashboards for Apache and the two systems to see if anything was overloaded.
What is the transaction rate -- you can get that from the Apache process dashboard?
Is Apache doing anything else other than forwarding the requests?
-- Graeme
Hi Graeme,
Looking at the process health, it seems it is fine.
Also I have noted there are fast requests indeed.
Transactions:
