Driven 2.0 Release Notes

version 2.0.5

New in this Release

Extended integration with other big data frameworks

Driven renders diagrams and reports metrics for data applications that are coded in Hive or with MapReduce, in addition to ongoing integration with Cascading. See the Driven Agent Getting Started for how to integrate your Hive and MapReduce applications with Driven.

Product terminology changed to reflect more application entities

To reflect broader application framework support, Driven 2.0 has renamed some user interface elements to be less Cascading-centric:

Flows are now called Units-of-Work
Taps are now called Resources

Status overview pages

There are two new landing pages instead of the one SHOW ALL page of previous Driven versions:

SHOW ALL APPS displays a Status View compilation of cluster-level application performance information that you are permitted to view. This page shows data for apps that your Driven teams monitor as well as apps of other teams that do not restrict discoverable metrics.
SHOW MY APPS is a Status View compilation of the applications that are associated with all your Driven teams. The data set that this page collects is the same as the SHOW ALL page of previous Driven versions

See Starting with an Overview of Monitored Applications.

Search functionality on app details pages

An app details page has a search box that can be filtered by specific Unit-of-Work (formerly Flow), Process, Statement, and Resource (formerly Tap) details. This search capability introduces the ability to set the search parameter to retrieve units-of-work that match specified metadata (ID, name, or tags) after you drill down to the application run that interests you.

The Unit-of-Work search filter is particularly helpful if you want to focus on particular aspects of an application execution that has processed numerous units of work. The directed acyclic graph (DAG) of an app details page can be difficult to inspect if the graph renders an excessive amount of activity.

See Searching App Details.

Enhanced Status View

The Status View has a new control and a different graphical display to improve discovery of relevant data:

A timeline selector grid shades intervals in proportion to application executions for a specific period. The visual cues of the shaded grid enable one-click access to time intervals with relevant application status data, instead of requiring the user to engage in a longer process of discovery.
The Status Summary graph (formerly the Status Timeline graph) is a line graph (rather than a bar graph) to provide a better visual indicator for you to compare application statuses between different points in time.
Both graphs display PENDING, STARTED, and RUNNING statuses as separate dimensions, instead of plotting these three states as one accumulated statistic.

See Status Views.

More controls on the display of application performance data

A user’s access to Status View information and application details can vary based on either server-wide or team-level configurations. New discoverability and visibility settings help users to focus on relevant data and enhance security of information about monitored applications.

See Sharing and Limiting Access to Application Information.

Change in pre-canned statistics on Application Views

The Time Between Failures statistical metric is no longer available.

Fixed Bugs

2.0.5 Plugin side configurations such as driven.protocol.slice.suppress are no longer being ignored.
2.0.4 Certain counter aliases were using the values of mapred and mapreduce counters when Hadoop 2 populated both. Now only using mapreduce.