The camunda Cockpit Statistics Plugin is a camunda BPM community extension, which provides a statistics plugin for the camunda Cockpit. It creates a set of charts helping you to understand what was and what is going on in your engine. At the moment a variety of charts is already available. As soon as you add the plugin, the appearance of camunda cockpit will be modified. On the front page you can now see a tabbed pane with information about all processes, user tasks, the timings and further analytics.
We are developing this plugin for the camunda cockpit, because several of our customers asked for it and we also missed this function in the cockpit. It can be very interesting to add an analytics tool like this to show the comparison between process instances. Now it is possible to provide these data to business departments, which control the implemented process, and can change wrong or long running processes in cooperation with the IT department.
The plugin was first presented at the camunda community day (http://blog.camunda.org/2014/09/camunda-bpm-community-day.html) by us (thanks to Eric) and at the BPMCon 2014 within an official presentation by Camunda. Thanks to Falko Menge who presented our statistics plugin on the BPMCon.
Now what can this plugin do for you? The process instances are shown in the tabbed menu. Three pie charts provide information about the running, ended and failed process instances. This statistic refers to all known process instances. The second tab describes all available running and ended user tasks with a pie chart. The timing tab shows a timeline, containing information about the start and end times of process instances and activities per process definition. The last tab, analytics, shows the durations of all process instances of one or more selected process definitions plotted on a timeline. It is also possible to draw a regression line, so one can see if there is a linear trend.
Further statistics are displayed for each process definition. These are reachable using the tab at the lower block. All statistics derive from the selected process definition. The statistic is separated into three charts: The first is a pie chart of running and ended user tasks per process definition. The second contains a pie chart of finished activities per process definition including information on avg, min, max duration and count. The last one shows the durations of the selected task.
To include the plugin into your cockpit you can either include it in your custom build on camunda’s plugin store or you build the cockpit on your own and deploy it to your server. Do not forget to customize build.properties in case you choose the latter option. The blog post http://blog.camunda.org/2014/08/cockpit-plugin-store.html describes the whole functionality of the new plugin store by camunda.
If you want to write your own plugin, with different features, then you can read these tutorials on the camunda website: http://docs.camunda.org/latest/real-life/how-to/#cockpit or http://docs.camunda.org/latest/guides/user-guide/#cockpit-plugins The tutorials contain everything you need to know in order to write an own cockpit plugin.
So what’s next? First, we want to fix a few performance issues in the chart showing start and end times of process instances per process definition and all available process instances. Furthermore, we want to improve the performance, test the plugin with real data and add charts visualizing process of different attributes like start / end time, duration, count or others based on the several existing resources. Regarding our overall architecture, we want to provide an easy way to integrate other data resources that may already exist in companies. Firstly, we are planning to integrate the elasticsearch plugin (https://github.com/camunda/camunda-bpm-elasticsearch) by Christian Lipphardt and we will evaluate the connection to the open source statistic framework R (http://rforge.net/Rserve/) for more advanced computation, in addition.
If you have an idea to modify or add something to the plugin, please leave a comment below. Or you can contribute to the project with a fork and a pull request to this repository https://github.com/camunda/camunda-cockpit-plugin-statistics on GitHub. Did you test the plugin and have experience how it works with real process data? Your feedback is most welcome to us to improve the implemented features.