SWS Conn Queues Test

In Web Server, acceptor threads on a listen socket accept connections and put them into a connection queue. Request processing threads in a thread pool then pick up connections from the queue and service the requests.

If the connection queue becomes full, the server will be forced to drop new connections, thus denying users access to the server. To avoid such an unpleasant outcome, it is imperative that the connection queues are monitored periodically.

This test keeps track of the changes to the size of the connection queues, so that administrators are proactively alerted to any abnormal increase in the queue size or an unusual delay while connections are being picked up for processing.

Target of the test : An Oracle iPlanet Web Server

Agent deploying the test : An internal agent

Outputs of the test : One set of results for each volume that is to be monitored.

Configurable parameters for the test
Parameter	Description
Test period	How often should the test be executed.
Host	The host for which the test is to be configured.
Port	The port number at which the specified Host listens to.
User	Specify the name of the administrator of the administration server to which the target web server is registered.
PasswdFile	The eG agent connects to the administration server to which the target web server is registered to extract the required metrics. To enable the eG agent to communicate with this administration server, you first need to create a password file containing the administrator password for accessing the server; this password file should be created in the install directory of the web server. Specify the name of this file in the PasswdFile text box; by default, this is admin.passwd.
AdminServer	Specify the IP address of the administration server to which the target web server is registered.
AdminPort	Indicate the port at which the administration server listens.
Config	In Web Server 7.0, all configurable elements of a server instance like web applications, configuration files, and search collection indexes are logically grouped and termed as a Configuration. To enable the test to locate the web server instance to be monitored, you need to mention the Configuration that applies to the instance. In the Config text box therefore, specify the name of the Configuration which the target web server instance uses.
Node	Node is a network resource, such as a server or a host. Instance refers to the environment of a web server daemon on a given node, including its configuration, log files and other runtime artifacts such as lock databases, caches and temporary files. To enable the test to identify the exact web server instance to be monitored, you need to specify the name of the Node on which the instance executes. The test will thus monitor only that web server instance that operates on the given Node and which uses the specified Config.
SSL	Set this flag to Yes if the administration server is SSL-enabled; if not, set it to No.
WebServerDir	Here, specify the Oracle iPlanet web server installation directory.

Measurements made by the test
Measurement	Description	Measurement Unit	Interpretation
Present Conns Queued	Indicates the number of connections currently in the queue.	Number
No. of connections queued	Indicates the total number of times a connection was queued during the last measurement period.	Number	This number includes newly-accepted connections and connections from the keep-alive system.
No. of connections overflowed	Indicates the number of connections that overflowed during the last measurement period.	Number	If the connection queue is full, then the acceptor threads will drop subsequent connections. The dropped connections represent the overflowed connections. Ideally, the value of this measure should be 0.
Maximum queue size	Indicates the maximum size of the connection queue.	Number	This number is: Maximum Queue Size = Thread Pool Queue Size +Maximum Threads+Keep-Alive Queue Size If the value of this measure is close to the limit, you can increase the maximum connection queue size to avoid dropping connections under heavy load.
Ticks spent	Indicates the amount of time that connections spent in the connection queue during last measurement period.	Secs
Average queuing delay	Indicates the average amount of time a connection spent in the connection queue during last measurement period.	Secs	This represents the delay between when a request connection is accepted by the server and when a request processing thread begins servicing the request. If the value of this measure is high, it could indicate a processing bottleneck on the server.
Peak queue size	Indicates the largest number of connections that have been in the queue simultaneously.	Number