Integration of the Deployment Visualization and Analysis

How to use

If you want to use the iobserve analysis with the integrated deployment visualization, follow the steps below.

Basically three components have to work together in this case. The replayer sends kieker records to the analysis. The analysis works on those and additional data (namely pcm models) and sends visualization elements in JSON to the deployment visualization. In the end the user can look at the analyzed system and its components in the browser.

Prerequisites

You need iobserve-ui-deployment (branch: jweg-iobserveIntegration) and iobserve-analysis (was formerly in branch: master-project-ss17). Follow the install instructions of those projects.

Execution

For the execution you need the deployment visualization, the iobserve-service and the replayer running in this order.

• You will find the information on how to start the deployment visualization in the instructions of iobserve-ui-deployment.
• For the replayer you need arguments as follows:
  • -i, --input path to a kieker logfile directory
  • -p, --port output port where the data is send to
  • -h, --host name or IP address of the host where the data is send to

In our case the port and the host refers to the iobserve-service.

• For the iobserve-service you have arguments as follows:
  • --help show usage information
  • -i, --input port number to listen for new connections of Kieker writers This has to match the output port from the replayer.
  • -p,--pcm path to a directory containing all PCM models
  • -pn4j, --pcmneo4j path to a directory containing Neo4j database with PCM model data
  • -o,--output hostname and port of the iobserve deployment visualization, e.g., localhost:9090.
  • -t,--think-time Variance of user groups for the clustering
  • -V,--variance-of-user-groups Variance of user groups for the clustering

Tasks overview and first ideas

1. Integration of the Deployment Visualization and Analysis

• A stage has to be created that pushes/saves the latest model from the knowledge repository to the database of the deployment visualization. The VisualizationUpdateStage (in analysis.service) is taken as a starting point and is developed further. In a first step VisualizationUpdateStage is split into two stages.
• Further the actual stages that handle deployments TDeployment, undeployments TUndeployment and allocations TAllocation should be modified and rearranged, such that the control flow is more obvious.

2. Extension of the Deployment Visualization

• The communication lines have to be normalized, such that different workloads can be distinguished easier.

3. Integration of Deployment Visualization and User-Behavior Visualization

Timetable

6.7.

• send an example query to iobserve-ui-deployment
• rearrange and modify TDeployment, TUndeployment and TAllocation
• implement VisualizationUpdateStage for TAllocation
• implement VisualizationUpdateStage for TDeployment
• implement VisualizationUpdateStage for TUnDeployment
• initialize visualization, send whole model to iobserve-ui-deployment
• test and document (javadoc, how to use), more information about testing stages in TeeTime

3.8.

• extend deployment visualization: visualize usergroups and performance, too
• test and document (iobserve-ui API docs)

7.9.

• connect deployment-visualization with user-behavior visualization
• test and document (iobserve-ui API docs)