Jump to: navigation, search

Difference between revisions of "SMILA/Project Concepts/Performance Measurement Task Summary"

(New page: == Eclipse TPTP Demonstration == This demonstration contains the following: - Extending Statistical Analysis launch configuration. - Building and monitoring Sample Statistical Agent. - Bui...)
 
 
Line 10: Line 10:
 
Statistical launch configuration type provides interface to execute particular agent with given options. It can be extended by building the Eclipse plugin that extends org.eclipse.hyades.perfmon.PerfmonPlugin and org.eclipse.hyades.perfmon.internal_agentExtension extension point. Plugin must provide Agent specific information such as Agent name and Process executable and options that are required to launch the agent (if needed).
 
Statistical launch configuration type provides interface to execute particular agent with given options. It can be extended by building the Eclipse plugin that extends org.eclipse.hyades.perfmon.PerfmonPlugin and org.eclipse.hyades.perfmon.internal_agentExtension extension point. Plugin must provide Agent specific information such as Agent name and Process executable and options that are required to launch the agent (if needed).
  
[[Image:test-agent-launch.png]]
+
[[Image:test-agent-launch.png|thumb]]
  
  
 
=== Sample Statistical Agent ===
 
=== Sample Statistical Agent ===
 
Sample Statistical Agent that has only one counter that is updated with random values is included in the demo. After agent is properly configured, it can be launched and monitored with Test Agent launch configuration:
 
Sample Statistical Agent that has only one counter that is updated with random values is included in the demo. After agent is properly configured, it can be launched and monitored with Test Agent launch configuration:
[[Image:sample-agent-monitoring.png]]
+
[[Image:sample-agent-monitoring.png|thumb]]
  
 
After the monitoring is stopped, profiling data can be saved and statcon file can be created. Statcon file will hold all of the settings for analysis of the collected statistical data:
 
After the monitoring is stopped, profiling data can be saved and statcon file can be created. Statcon file will hold all of the settings for analysis of the collected statistical data:
  
[[Image:statcon-file.png]]
+
[[Image:statcon-file.png|thumb]]
 
+
 
+
== Eclipse TPTP Demonstration ==
+
This demonstration contains the following:
+
- Extending Statistical Analysis launch configuration.
+
- Building and monitoring Sample Statistical Agent.
+
- Building and monitoring sample JMX agent.
+
 
+
 
+
===Adding custom Statistical Analysis launch configuration ===
+
Statistical launch configuration allows to launch particular data collection Statistical Agents. Statistical Agent is and XML message-based agent for real-time monitoring of operating systems, web servers, database servers and application servers. The following agents are provided with TPTP distribution: Windows System Performance Data (Perfmon), Linux System Perfmon Data, JBoss Application Server JMX Data, JOnAS Application Server, JMX Data, Apache/IHS server-status Data, MySQL Database Table and Variable Data
+
Statistical launch configuration type provides interface to execute particular agent with given options. It can be extended by building the Eclipse plugin that extends org.eclipse.hyades.perfmon.PerfmonPlugin and org.eclipse.hyades.perfmon.internal_agentExtension extension point. Plugin must provide Agent specific information such as Agent name and Process executable and options that are required to launch the agent (if needed).
+
[[Image:test-agent-launch.png]]
+
 
+
 
+
 
+
=== Sample Statistical Agent ===
+
Sample Statistical Agent that has only one counter that is updated with random values is included in the demo. After agent is properly configured, it can be launched and monitored with Test Agent launch configuration:
+
 
+
[[Image:sample-agent-monitoring.png]]
+
 
+
After the monitoring is stopped, profiling data can be saved and statcon file can be created. Statcon file will hold all of the settings for analysis of the collected statistical data:
+
 
+
[[Image:statcon-file.png]]
+
 
+
  
  
 
=== Sample JMX Agent ===
 
=== Sample JMX Agent ===
 
TPTP contribution includes jmxagent that can be extended to collect JMX data from any JMX server. To do this you have to implement org.eclipse.hyades.perfmon.agents.jmx.ascode.ServerCommunicationInterface and org.eclipse.hyades.perfmon.agents.jmx.ascode.ServerStatInterface interfaces and provide XML configuration for the statistical information that will be collected. TPTP distribution includes jmxagent implementations for JBoss and JOnAS Application Servers. This demonstration includes sample Tomcat JMX Data collection agent implementation. The same Test Agent Statistical launch configuration is used to start monitoring:
 
TPTP contribution includes jmxagent that can be extended to collect JMX data from any JMX server. To do this you have to implement org.eclipse.hyades.perfmon.agents.jmx.ascode.ServerCommunicationInterface and org.eclipse.hyades.perfmon.agents.jmx.ascode.ServerStatInterface interfaces and provide XML configuration for the statistical information that will be collected. TPTP distribution includes jmxagent implementations for JBoss and JOnAS Application Servers. This demonstration includes sample Tomcat JMX Data collection agent implementation. The same Test Agent Statistical launch configuration is used to start monitoring:
[[Image:tomcat-jmx-statistical.png]]
+
[[Image:tomcat-jmx-statistical.png|thumb]]
  
  
Line 55: Line 30:
 
Following diagram represents the architecture of the JMX Application Data Collection with TPTP:
 
Following diagram represents the architecture of the JMX Application Data Collection with TPTP:
  
[[Image:diagram.png]]
+
[[Image:diagram.png|thumb]]
  
 
== Further tasks overview ==
 
== Further tasks overview ==
Line 68: Line 43:
 
### simple to use in existing Java applications
 
### simple to use in existing Java applications
 
# Analyze how collected Statistical Performance Traces can be combined together. (Traces from Operationg System Performance Information and from the Eccence Performance Counters)
 
# Analyze how collected Statistical Performance Traces can be combined together. (Traces from Operationg System Performance Information and from the Eccence Performance Counters)
# Probably also create Logging Agent in addition to Statistical to create CBE Logs directly from JMX data. The monitored Values should be savable (used of CBE possible?) anad graphs information should be loadable/storable
 
# Documentation: Installation and Use of the API
 
 
 
[[Category:SMILA]]
 
=== Sample JMX Agent ===
 
TPTP contribution includes jmxagent that can be extended to collect JMX data from any JMX server. To do this you have to implement org.eclipse.hyades.perfmon.agents.jmx.ascode.ServerCommunicationInterface and org.eclipse.hyades.perfmon.agents.jmx.ascode.ServerStatInterface interfaces and provide XML configuration for the statistical information that will be collected. TPTP distribution includes jmxagent implementations for JBoss and JOnAS Application Servers. This demonstration includes sample Tomcat JMX Data collection agent implementation. The same Test Agent Statistical launch configuration is used to start monitoring:
 
[[Image:tomcat-jmx-statistical.png]]
 
 
 
=== Software overview diagram ===
 
Following diagram represents the architecture of the JMX Application Data Collection with TPTP:
 
[[Image:diagram.png]]
 
 
== Further tasks overview ==
 
# Build the api framework for performance measurement based on JMX for integration in Eccenca project (similar to .NET performance counters api?). Create API Framework for Eccenca Performance Counters
 
## neccessary Performance Counter
 
### Performance Counter
 
### Throughput Values
 
### Status Notification/Values
 
### transaction monitoring(like arm method calls?)
 
## Framework should be
 
### reusable
 
### simple to use in existing Java applications
 
# Analyze how collected Statistical Performance Traces can be combined together.(Traces from Operationg System Performance Information and from the Eccence Performance Counters)
 
 
# Probably also create Logging Agent in addition to Statistical to create CBE Logs directly from JMX data. The monitored Values should be savable (used of CBE possible?) anad graphs information should be loadable/storable
 
# Probably also create Logging Agent in addition to Statistical to create CBE Logs directly from JMX data. The monitored Values should be savable (used of CBE possible?) anad graphs information should be loadable/storable
 
# Documentation: Installation and Use of the API
 
# Documentation: Installation and Use of the API

Latest revision as of 11:00, 15 August 2008

Eclipse TPTP Demonstration

This demonstration contains the following: - Extending Statistical Analysis launch configuration. - Building and monitoring Sample Statistical Agent. - Building and monitoring sample JMX agent.


Adding custom Statistical Analysis launch configuration

Statistical launch configuration allows to launch particular data collection Statistical Agents. Statistical Agent is and XML message-based agent for real-time monitoring of operating systems, web servers, database servers and application servers. The following agents are provided with TPTP distribution: Windows System Performance Data (Perfmon), Linux System Perfmon Data, JBoss Application Server JMX Data, JOnAS Application Server, JMX Data, Apache/IHS server-status Data, MySQL Database Table and Variable Data Statistical launch configuration type provides interface to execute particular agent with given options. It can be extended by building the Eclipse plugin that extends org.eclipse.hyades.perfmon.PerfmonPlugin and org.eclipse.hyades.perfmon.internal_agentExtension extension point. Plugin must provide Agent specific information such as Agent name and Process executable and options that are required to launch the agent (if needed).

Test-agent-launch.png


Sample Statistical Agent

Sample Statistical Agent that has only one counter that is updated with random values is included in the demo. After agent is properly configured, it can be launched and monitored with Test Agent launch configuration:

Sample-agent-monitoring.png

After the monitoring is stopped, profiling data can be saved and statcon file can be created. Statcon file will hold all of the settings for analysis of the collected statistical data:

Statcon-file.png


Sample JMX Agent

TPTP contribution includes jmxagent that can be extended to collect JMX data from any JMX server. To do this you have to implement org.eclipse.hyades.perfmon.agents.jmx.ascode.ServerCommunicationInterface and org.eclipse.hyades.perfmon.agents.jmx.ascode.ServerStatInterface interfaces and provide XML configuration for the statistical information that will be collected. TPTP distribution includes jmxagent implementations for JBoss and JOnAS Application Servers. This demonstration includes sample Tomcat JMX Data collection agent implementation. The same Test Agent Statistical launch configuration is used to start monitoring:

Tomcat-jmx-statistical.png


Software overview diagram

Following diagram represents the architecture of the JMX Application Data Collection with TPTP:

Diagram.png

Further tasks overview

  1. Build the api framework for performance measurement based on JMX for integration in Eccenca project (similar to .NET performance counters api?). Create API Framework for Eccenca Performance Counters
    1. neccessary Performance Counter
      1. Performance Counter
      2. Throughput Values
      3. Status Notification/Values
      4. transaction monitoring(like arm method calls?)
    2. Framework should be
      1. reusable
      2. simple to use in existing Java applications
  2. Analyze how collected Statistical Performance Traces can be combined together. (Traces from Operationg System Performance Information and from the Eccence Performance Counters)
  3. Probably also create Logging Agent in addition to Statistical to create CBE Logs directly from JMX data. The monitored Values should be savable (used of CBE possible?) anad graphs information should be loadable/storable
  4. Documentation: Installation and Use of the API