ICAP Server Status

You can monitor the ICAP Server and content scanning engine status, and restart stuck engines by using ICAP Server Status.

From the ICAP Gateway main page, click Health > ICAP Server Status

The page displays:

• ICAP Server CPU usage
• A list of content scanning engines that process outbound traffic, and their status
• A list of content scanning engines that process inbound traffic, and their status

Tell me about...

• The information that is displayed

  Icon	State
  	Engine idle and waiting for job
  	Engine processing
  	Engine restarting

  Other information displayed:

  Item	Description
  ICAP Server CPU Usage (%)	Percentage of time where the CPU is processing instruction as opposed, for example, waiting for input and output operations.
  Queue Size	The number of jobs that are waiting to be processed.
  Engine PID	PID, or process identifier, is a number that is used to uniquely identify a process.
  CPU Usage	Percentage of time where the CPU for the given PID is processing instruction as opposed, for example, waiting for input and output operations.
  Client	The username and IP address of the client machine that has generated the job.
  Job Size	The size of the job that is being processed, for example, the size of a file being downloaded.
  Elapsed Scan Time	The total processing time for the given PID on this job.
  Node Count	The number of nodes that have been processed so far for the request.
  Node Scan Time	The time taken by processing the last node used in the request.
  URL	The URL being processed by the engine.

• Why content scanning engines get stuck

  With the volume and variation of data that can be streamed through your Gateway it is possible that a content scanning engine will encounter data that it cannot understand and process successfully. In most cases, engines will complete processing and move onto the next job. However, in a minority of cases, an engine might not recover from trying to process data it does not understand. When this happens, the engine becomes stuck and does not become available for other jobs until it is restarted. The Watchdog service automatically restarts the engine following a timeout. You can also manually enforce this operation using the ICAP Server Status page.

• What happens to an engine when I restart it

  When a restart request is made, the specified content scanning engine will be terminated, and another engine will be started to take its place. Any existing transactions for the engine being restarted will be terminated.

  An icon will display in the user interface against the engine being restarted, indicating progress of the restart. The restart event will be logged in the Web Proxy Event log.

  Depending on what an engine is processing when it is restarted, users might receive an error message in the browser. For example, if a file download is terminated, an HTTP 500 error message might be encountered.

How do I...

• Use the information to determine if an engine is stuck?

  The information provided on this page can provide evidence of an engine that is stuck processing a transaction. For example, an engine with an unchanged node count, high node scan time, and a CPU usage that is at 0 or 100% consistently, is strong indicator than an engine is stuck.

  Stuck engines might not always be due to problems with the server. For example, small queues or several idle engines could indicate a different problem.

  As engines are displayed by processing time, any engines that are stuck are likely to be at the top of the list.

• Check a URL that is being processed by an engine?

  If you think an engine is having difficultly processing a request, you can take a copy the URL for further investigation. To do this:

  1. Click on the engine in the table view.
  2. Click Show URL. A dialog box with the URL will be displayed.
  3. Select and copy the URL from the dialog box.
  4. Click Close.

  You should use a different Gateway to check the URL.

• Restart an engine?

  1. Click on the engine in the table view.
  2. Click Restart Engine. A dialog box is displayed stating that all existing transactions will be terminated for the selected engine.
  3. Click Restart. A dialog box is displayed stating that a request to restart the proxy engine has been sent.

  After the restart a dialog box will display the result of the request.

See also...

• Monitoring and Managing ICAP Server Performance for an overview of proxy monitoring tools
• ICAP Watchdogs for automatic proxy monitoring