Video: Health management for Liberty collectives

The following transcript is for the "Health management for Liberty collectives" video, which demonstrates the benefits of using health management in your Liberty environment. This transcript is the video storyboard. Audio describes narration and captions. Onscreen Action describes the content shown in the video.

Video Health management for Liberty collectives

Table 1. Introduction and overview. Show title and introduce health management.
Scene Audio Onscreen Action
1 Do your applications use too much memory, have memory leaks or slow response time? Health management in Liberty can help find and mitigate problems in your applications as they occur. Show title Health management for Liberty collectives.
2 Health management in a Liberty collective can prevent the disruption of service by detecting common problems and generating diagnostic actions based on configured health policies. Show words:

Prevent disruption of service

Use policies to detect app health concerns:
  • Excessive memory usage
  • Memory leaks
  • Excessive request timeout
  • Excessive average response time
3 You can target health policies at individual servers, all servers in a cluster, or all servers on a host. Show diagram of server clusters with hosts and highlight each section.
4 To diagnose and manage errors, health policies can capture heap or thread dumps, restart servers, and force servers to enter maintenance mode. Show words:
Health policies can:
  • Capture heap and thread dumps
  • Restart servers
  • Force servers into maintenance mode
Table 2. Feature and JVM configuration for health management.. Show the process for configuring health management.
Scene Audio Onscreen Action
5 You can add support for health management to your Liberty collectives through the healthManager-1.0 and healthAnalyzer-1.0 features. After you set up a collective, add the healthManager-1.0 feature to a collective controller. If you want to monitor excessive memory usage or memory leaks, add the healthAnalyzer-1.0 feature to a collective member server. Show code for adding healthManager-1.0 to a collective controller server.xml file and adding healthAnalyzer-1.0 to a collective member server.xml file.
Table 3. Health policy configuration for health management.. Show an example of a health policy configuration for health management.
Scene Audio Onscreen Action
6 To configure a health policy, add a healthPolicy element to the collective controller server.xml file. You can configure multiple policies by adding additional healthPolicy elements.

Each health policy has one or more targets, one condition, and one or more actions.

The target can be a cluster, a host, or a server. This healthPolicy element targets a single server.

The condition that triggers corrective action is excessive average response time of at least 10 seconds. When this condition exists, the affected server is put into maintenance mode. If you specify multiple actions, the actions run in the order that they are specified in the policy.

Show a controller server.xml file with the example healthPolicy element highlighted.
<healthPolicy id="myPolicy">
   <server hostName="Host" wlpUsrDirectory="ibm/liberty/wlp" serverName="server1"/>
   <excessiveResponseTime responseTime="10s"/>
   <action action="enterMaintenanceMode"/>
</healthPolicy>
7 Let's take a look at this policy in action.

The healthManager-1.0 feature is enabled in a collective controller that is configured for dynamic routing and auto-scaling.

Show a controller server.xml file with the example healthPolicy element highlighted.
<featureManager>
   <feature>jsp-2.2</feature>
   <feature>collectiveController-1.0</feature>
   <feature>dynamicRouting-1.0</feature>
   <feature>scalingController-1.0</feature>
   <feature>healthManager-1.0</feature>
</featureManager>

...

<healthPolicy id="myPolicy">
   <server hostName="Host" wlpUsrDirectory="ibm/liberty/wlp" serverName="server1"/>
   <excessiveResponseTime responseTime="10s"/>
   <action action="enterMaintenanceMode"/>
</healthPolicy>
8 The Intelligent Management IBM® HTTP Server plug-in detects average response times of over 10 seconds for Server 1. In accordance with the health policy, the controller forces Server 1 into maintenance mode, and IBM HTTP Server stops routing new requests to that server.

As Server 1 goes into maintenance mode, the scaling controller simultaneously detects that fewer than two servers are running and starts Server 3. IBM HTTP Server begins routing requests to that server.

Show animation of a server being put into maintenance mode, another server starting, and traffic being routed to the newly started server.
Table 4. Conclusion. Show where to find more information about health management.
Scene Audio Onscreen Action
9 For more information about health management and Liberty collectives, see the Knowledge Center documentation or WASdev.net. Show information on documentation:
IBM Knowledge Center
http://www-01.ibm.com/support/knowledgecenter/
WASdev.net
http://developer.ibm.com/wasdev

For more information, see Configuring health management for Liberty.


Icon that indicates the type of topic Concept topic

File name: video_transcript_hm.html