Sun Java System Web Server 7.0 Update 7 Administrator's Configuration File Reference

init-request-limits

The init-request-limits function works with the obj.conf function check-request-limits to monitor incoming requests with a given attribute. check-request-limits maintains a table of monitored values. intit-request-limits purges existing entries in that table according to the timeout. This function is not required unless you want to override the default value for the purge timeout in check-request-limits. For more information, see check-request-limits. The default is 300 seconds (five minutes). This function is applicable in Init-class directives.

Parameters

The following table describes the init-request-limits parameter.

Table 5–4 init-request-limits Parameter

Parameter 

Description 

timeout

(Optional) Sets the time in seconds after which to purge entries tracked by check-request-limits. The default is 300 seconds (five minutes).

An optimal value for timeout depends not only on your performance and memory requirements but also on the check-request-limits rules you are using. When using rules containing, for example, monitor="$ip" on a busy public web site, new buckets are created and kept for every client IP accessing the server. Because this setting potentially creates a very large number of buckets, the expiration should be short enough that unused entries are purged in a reasonable time.

However, to avoid removing and re-creating buckets for the same client, do not set a timeout that is shorter than the typical or expected client session.

If you do not use any dynamic bucket names (that is, if all monitored values and bucket are fixed strings instead of variables, or you never specify monitor or bucket parameters at all) there are only a fixed number of buckets. In that case, you can disable expiration entirely by setting the timeout to zero.

Example

Init fn="init-request-limits" timeout="120"