Open with table of contents

ICAP Watchdogs

You can automatically monitor and manage proxies and their engines by using:

• The Content Scanning Engine Watchdog
• The ICAP Server Watchdog

The ICAP Server Watchdog is designed to detect if the ICAP Server is unresponsive and restart it automatically, while the Content Scanning Engine Watchdog is designed to detect content scanning engine failure and provide automatic restart. The different modes of operation for both Watchdogs mean that the feature can be switched on or off and can be configured to provide notifications.

Tell me about...

• The modes of operation for each Watchdog

  Mode	Description
  Off	The Watchdog is disabled. No alarms are raised, and no restart of an unresponsive proxy or stuck engines will occur.
  Warning	An alert will be raised indicating that a proxy is unresponsive or an engine is stuck. No automatic restart will occur.
  Automatic	Unresponsive proxies and stuck engines will be restarted automatically. An alarm will be raised indicating that a problem occurred and it has been resolved. Data for the event will be logged to aid diagnostics.

• What happens if I switch the Watchdog features off

  The Gateway will function normally if these features are switched off. However, unless you are monitoring the server manually, you will not be aware of any problem engine processes with possible performance degradation. If you choose not to have automatic engine restart, setting the mode to Warning will allow you to act on an alarm and manually check engine processes by using ICAP Server Status.

• What happens if the Watchdog is in Warning mode and I ignore alerts

  You might experience performance degradation in the service. It is recommended that any alerts raised are investigated using ICAP Server Status.

• The criteria the Watchdogs use to restart the service and if I can configure it

  The Content Scanning Engine Watchdog examines the length of time an engine has spent processing a node. If this time exceeds a given threshold, the Watchdog determines that the engine is in a stuck status.

  You can configure the threshold times for three ranges of job sizes:

  Job size	Default threshold timeout (minutes)
  0-100MB	10
  100MB-1GB	20
  1GB+	30

  The ICAP Server Watchdog uses a heartbeat to determine if the server is still responsive; a file is written to at a regular interval and if the Watchdog detects that the file has not been updated within the defined time, a notification and, if configured, a restart will occur.

  We recommend that you change these configuration settings only under the guidance of Clearswift support personnel.

How do I...

• Configure the Watchdogs?

  1. From the Home page, click System > ICAP Server Monitoring.

     The ICAP Server Monitoring page is displayed.

  2. Click the Watchdog tab.

     The current configurations for both Watchdogs are displayed in separate panels.

  3. For the Watchdog you want to change, click Click here to change these settings in the appropriate panel.
  4. Select the options you want and click Save.

See also...

• Monitoring and Managing ICAP Server Performance for an overview of proxy monitoring tools
• ICAP Server Monitoring for manual monitoring and restart of proxy engines