Administrator toolkit

Modify solution and system behavior

Use the solutionManager and propertyManager scripts to adjust default values and add new properties. For more information, see Properties.

You set a system property on a single server of a server topology; system property updates are pushed to all servers.

Manage system memory

Monitor key metrics to ensure that your runtime servers and solutions are working well.

Monitor the Java™ virtual machine (JVM) heap usage remotely by using the MBeans supplied by the JVM.

You can also use JConsole to get information on performance and resource consumption both on a local or remote computer. Connect to a JVM and then browse the JMX MBeans to see the wide variety of metrics available. The following batch file launches JConsole and connects to a remote JVM over HTTPS:

REM service:jmx:rest://hostname:9443/IBMJMXConnectorREST
REM username is tester
REM password is tester
REM you can set -J-Djavax.net.debug=ssl to help debug
REM **NOTE** SSL host name verification is disabled.

C:/IBM/ODMInsights/jdk/bin/jconsole 
-J-Djava.class.path=C:/IBM/ODMInsights/jdk/lib/jconsole.jar;
C:/IBM/ODMInsights/jdk/lib/tools.jar;
C:/IBM/ODMInsights/runtime/wlp/clients/restConnector.jar 
-J-Dcom.ibm.ws.jmx.connector.client.disableURLHostnameVerification=true 
-J-Djavax.net.ssl.trustStore=C:/IBM/Solutions/Config/key.jks 
-J-Djavax.net.ssl.trustStorePassword=tester 
-J-Djavax.net.ssl.trustStoreType=jks

If the heap usage exceeds some percentage of available heap, alarms can be generated.

Do not expect the system to automatically page entities in/out of memory.

Update entities

When you use the REST API to update an entity, agents are not notified that the entity is updated. If this lack of notification is an issue in your solution, submit an event to update your entities rather than using the REST API.

If the events are submitted by the gateway API, the events are routed to the correct partition and processed immediately. If the events are submitted through HTTP or JMS, the processing passes through the solution connectivity, which can be scaled independently of agent processing.

Protect access to entities

Configure grid security in a production server that is running in an unprotected network.

Do not keep test users and authorization roles in the configuration of your production server; at a minimum encode or encrypt the passwords.

Deploy solutions

In production, use the solutionManager script to deploy a solution.

In a production topology, use bash scripts over Secure Shell (SSH) to clean up the servers, start/stop the servers, deploy, and retrieve logs.

Pass a script and have it run ephemerally. For example:

cat script.sh | ssh user@remote_server /bin/bash

Or:

ssh user@remote_server "$(< localfile)"

In production, do not deploy a solution from Insight Designer.

Monitor your operations

Decision Server Insights provides comprehensive and flexible logging, and utilities to monitor the performance of Insight Server. Consider the following practices:

Centralize logging

Centralized logging provides an easy way to view and search your logs, which are normally stored locally on individual servers across a topology. Aside from the convenience of not having to log in to individual servers to read logs, centralized logging also helps to identify issues that span multiple servers by correlating the logs and metrics during a specific time frame. In a test environment, you can also use Insight Monitor, or install IBM® Log Analytics Server.

Choose which logs to enable

Gather the important logs from all of your servers, and set a time to retain the logs before they are discarded.

Use log filters

Filtering logs improves your ability to query, analyze, and graph the data in meaningful ways.

Synchronize your server clocks

Ensure that the clocks of your servers are synchronized and set to the same time zone, so that your log timeline is accurate across your topology.

Use the propertyManager script to enable system performance monitoring and capture warnings about garbage collection and available heap. The properties are automatically set on all servers when you run the script on one runtime server.

You can use the SystemMonitorMXBean API to retrieve warning messages collected by system monitoring. You can also reset the logging cache, and get notified whenever a warning message is added to the cache. The MBean returns only the warning messages of the runtime server to which the connection is established.

Create operational alerts to trigger an action, such as running a script or sending a notification, when a service or server goes down, or if a certain resource (CPU, memory, or storage) is over-utilized. These notifications help you to react to any issues as soon as they occur, which can help minimize or prevent the downtime of your application.

Build dashboards to improve operational efficiency. Some metrics are best viewed on a dashboard. IBM Log Analysis Server has a built-in index, search, and dashboard capability.

Use the pmiManager guardian utility to set conditions on the Insight Server MBeans. IBM Tivoli® Monitoring can also be used to set up alert conditions and alarms.

In production, do not enable every warning message and monitoring tool you can, unless execution speed is not an issue.

Monitoring is not always required for a production server, but the need for monitoring increases as a production environment grows in size and complexity. It provides an easy way to track your critical resources, and improve the planning and maintenance of your environment.

Test driver

Use test driver to set up and automate complex solution testing, and to detect server problems.

For more information, see TestDriver and Receiving and storing debug events and Creating automated test cases.

• Load entities into the data grid then submit test events.
• Debug a server problem by connecting a test driver instance and capturing events.

DELETE

Use DELETE to remove a business entity from a solution on an Insight Server.

For more information, see Deleting entities.

• Delete a single business entity.
• Delete all business entities of a specified type.
• Delete all entities in a solution.

DELETE state

For more information, see the REST API.

GET

Use GET to retrieve information about business entities in a solution.

For more information, see Listing entities.

• List the business entities in all deployed solutions.
• List the business entities in a specified solution.
• List all business entities that are associated with a specified solution.
• Display a single entity by key value.

POST

Use POST to update business entities in a solution. For example, use a POST request to change a flight number or customer account number.

For more information, see Creating entities.

• Update fields or data in a single business entity.
• Update fields or data in a set of business entities.

PUT

Use PUT to deliver business entities to a deployed solution, at the location indicated by the URI. You use the PUT request for delivering new business entities, and for delivering business entities that are updated with new fields or deleted fields. For more information, see Updating entity data.

• Deliver a newly created business entity to a deployed solution.
• Deliver an updated business entity to a deployed solution.

GET

Use GET to retrieve information about global aggregates in a solution.

For more information, see the REST API.

• Get information on a single global aggregate.
• Get information on a all global aggregates in a solution.

GET

Use GET to retrieve information about shared aggregates in a solution.

For more information, see the REST requests.

• Get information on a shared aggregate at a time point by using the when parameter.

connectivityManager

Use connectivityManager to manage the connectivity assets in a solution.

For more information, see Managing solution connectivity by running connectivityManager.

• Deploy solution connectivity.
• Generate solution inbound connectivity.
• Validate solution connectivity.
• Activate solution connectivity.
• Deactivate solution connectivity.
• Undeploy solution connectivity.

dataLoadManager

Use dataLoadManager to recover and restore configuration information and data that is persisted to a backup database.

For more information, see Restoring data after a system shutdown by running dataLoadManager.

• Load persisted data back onto the system.
• Restart the system manually after you load persisted data back onto the system.
• Load persisted data back onto the system, and restart the system automatically.
• Check the progress of a persisted data reload.

jobManager

Use jobManager to manage and obtain information about analytics jobs.

For more information, see Managing global aggregates.

• List all jobs, active jobs, or queued jobs.
• Display information about a job.
• Start a job.
• Update a job schedule.
• Stop a job.

outboundBufferManager

Use outboundBufferManager to manage the outbound connectivity (outbound endpoints), and the outbound events that are pending delivery, for a solution.

For more information, see Managing outbound endpoints by running outboundBufferManager.

• List all outbound endpoints, and outbound event counts, in a solution.
• Clear (discard) all pending events from the buffer of an outbound endpoint in a solution.

pmiManager

Use pmiManager to generate statistics about a running server. For more information, see pmiManager

• Generate statistics that include the number of times an agent or event type is invoked, average time an agent or event time is active, and the number of times an agent or event type is invoked per second.
• Set up a guardian utility to trigger an alert and detect anomalies when a server reaches specified thresholds.

propertyManager

Use propertyManager to manage server properties.

For more information, see Modifying server behavior by setting properties.

• Set server properties.
• Retrieve server properties.
• Modify server properties.

serverManager

Use serverManager to pause and shutdown server processes.

For more information, see Creating, starting, and stopping servers.

• Use isonline to determine whether a server is running.
• Use shutdown to pause and shutdown server processes and components in the correct order.
• Run suspendbalancing and resumebalancing to suspend and resume partition balancing across servers in a multiple-server production topology.

solutionManager

Use solutionManager to manage deployed solutions.

For more information, see Managing deployed solutions by running solutionManager.

• List all active servers.
• List all deployed solutions.
• Display the current solution version.
• Revert the solution to a previous version.
• Set or modify user authentication properties.
• Override the solution properties.

adminCenter-1.0 on a catalog server

monitor-1.0 on runtime servers

Use Insight Monitor to view event rates and available memory for all of the runtime servers.

Display metrics on each runtime server:

• Events: counts, rates, distribution
• Memory consumption
• CPU usage

console.log

Use console.log to view standard process output and error messages from the runtime environment, for run and debug actions. Messages are redirected to the server_name/logs/console.log file when you start the server.

For more information, see Trace and logging in the WebSphere Application Server, Liberty core documentation.

CWMBD0060I: Solution MyCreditCardSolution-0.0 ready

View other standard process output and error messages.

messages.log

Use messsages.log to view operational messages and information that is generated by the system.

For more information, see Trace and logging in the WebSphere Application Server, Liberty core documentation.

• View INFO, AUDIT, WARNING, ERROR, and FAILURE messages.
• View time stamps and the IDs of message threads.

trace.log

Use trace.log to gather and view debug information.

For more information, see Trace and logging in the WebSphere Application Server, Liberty core documentation.

Gather and view debug information that is obtained by basic, enhanced, or advanced tracing.