ICAP Server Performance

There might be occasions when you encounter performance issues with your ICAP Gateway. The traffic inspected by the ICAP Gateway depends on the policy configured in your existing ICAP client. Changes to policy and depth of inspection in the ICAP Gateway can impact performance.

This can be to due to a number of causes and can include:

• High volumes of web traffic
• Excessive number of concurrent ICAP connections
• ICAP Gateway processing large content.

Your ICAP Gateway has a number of features that allow you monitor and manage performance:

• The ICAP Server Monitoring page allows you to configure logging files.
• The Watchdog features enable you identify issues and resolve them automatically or, if preferred, generate warning messages to prompt for manual investigation.
• The features on the ICAP Server Status tab on the Health page allow you to monitor the activity of the content scanning engines, and restart problem engines as required.

Tell me about...

• The ICAP Server Monitoring Page

  The ICAP Server Monitoring page offers a number of features that allow you to investigate issues by means of logging and diagnostic reports:

  • ICAP Server Log Level

    You can use this feature to configure the minimum level of detail that is included in Gateway logs. You can also include specialized logging features such as Event Signaling and Engine Bypass (on request from a Clearswift Support Engineer).

  • ICAP Server Request Logging

    ICAP Server Request Logging allows you to log all requests and responses processed by the ICAP Server including data exchanged between the client and the proxy and the proxy and the server. Use this logging feature to investigate problems that occur when accessing a particular URL.

  The ICAP Server Monitoring page also contains the user interface for the Content Scanning Engine and ICAP Server Watchdog. Due to the nature of these features, they are described separately.

• The Watchdogs

  There are two Watchdogs.

  • The Content Scanning Engine Watchdog monitors content scanning engines, and will generate alarms and restart problem engines as appropriate.
  • The ICAP Server Watchdog monitors the proxy to ensure it is active, and will generate alarms and restart the proxy as appropriate.

  Both Watchdogs can be configured to work in different modes of operation. You can:

  • Choose to detect if there is a problem with engines or the ICAP Server automatically, thereby allowing you to ‘switch off’ the feature
  • On detection of a problem engine or inactive ICAP Server, raise an alarm
  • On detection of a problem engine or inactive ICAP Server, restart and raise an alarm
• ICAP Server Status

  ICAP Server Status allows you to check the status of content scanning engines by means of a User Interface in the ICAP Gateway. It is useful in general monitoring of your ICAP Server and in helping with problem diagnosis. For example, in the situation where an engine is stuck due to the processing of a particular URL, you can identify the URL and client for further investigation. In addition, the feature provides a mechanism for recovering content scanning engines that are stuck by allowing you to selectively restart them.

• The best method to adopt to monitor the ICAP Server

  If you want an immediate view and want to be able to restart individual proxy engines, you should use ICAP Server Status and the ICAP Server Watchdog features. You can only restart the content scanning engines by using ICAP Server Status.

  ICAP Server Status and Watchdogs

  By combining ICAP Server Status and the ICAP Server Watchdog features, you are able to monitor and manage both proxy and proxy content scanning engine processes.

  ICAP Server Status provides a detailed status of the ICAP Server and content scanning engines, and you are able to restart engines from this page. However, if you only use this method, you will need to monitor the status on a regular basis as problem engines will not be started automatically.

  The Content Scanning Engine Watchdog allows automatic identification and resolution of engines that have failed to process specific content. If the Watchdog has been configured to provide a warning only, you can use ICAP Server Status to investigate the issue and manually restart the engine if appropriate.

  Even if the Content Scanning Engine Watchdog is configured to restart engines automatically, you can resolve issues with problem engines sooner by performing a manual restart from ICAP Server Status.

  The ICAP Server Watchdog is similar to the Content Scanning Engine Watchdog in terms of the actions performed when the Watchdog has detected that the proxy is not active. That is, off (perform no action), generate alarm only, or generate alarm and automatic restart.

  Note that all watchdog activity is logged.

  We recommend that you configure both watchdogs to detect and resolve problems automatically. Use ICAP Server Status for when you want to check engine processes, or manually resolve reported problems.

How do I...

• Work with the features on the ICAP Server Monitoring Page?

  We recommend that you use these features only under the guidance of Clearswift support personnel.

  For further detail on these features, select the appropriate link:

  • Configure the level of log detail with ICAP Server Diagnostic Log Level
  • Log and then analyze requests and responses with ICAP Server ICAP Server Request Logging
  • Work with the Watchdogs
• Work with the Watchdogs?

  The user interface for the watchdogs is found on the Watchdog tab on the ICAP Server Monitoring page.

  For further detail on this page, including steps on how to configure the feature, see the ICAP Watchdogs topic.

• Work with ICAP Server Status?

  The User Interface for ICAP Server Status is found on the ICAP Server Status tab on the Health home page.

  For further detail about this page, including working with suspected problem URLs, and restarting engines, see the ICAP Server Status topic.

See also...

• Monitoring and Managing your Gateway for an overview of system performance tools