ElectricInsight — Software build analytics
ElectricInsight® is the only tool to visually depict the structure of a software build, down to the file level, empowering build managers to pinpoint performance problems or conflicts in a build. An add-on to ElectricAccelerator, ElectricInsight mines the information produced by ElectricAccelerator to provide an easy-to-understand, graphical representation of the build structure for performance analysis.
It provides detailed information and reports on each job on each node of the build cluster, for at-a-glance diagnostics. And it can predict and model how build times would be impacted by adding nodes to the ElectricAccelerator cluster, to help guide hardware investment decisions.
With ElectricInsight, instead of manually pouring over tens of thousands of lines in a build output file, error detection and performance tuning can now be done in seconds. Your teams can dramatically reduce time and costs associated with manual build maintenance tasks.
What can you learn about a software build using ElectricInsight?
- Are there any unnecessary serializations that are slowing the build down?
- Are there gaps where agents are not busy (and can I gain greater parallelism if I fix that)?
- Which are the longest job chains?
- Which files are being modified (created, updated, deleted) multiple times?
- Where exactly are the dependencies?
- If I added more agents to my cluster, how fast could it go?
View fine-grained, file-level detail on each build job to quickly perform diagnostics and improve build performance.
Click to view »
Clear reports identify job count, job time by length, longest jobs, and longest serial chains.
Click to view »
Model speed improvements that could be gained by adding additional hardware to the build cluster.
Click to view »ElectricInsight — Features and Benefits
| Visualize the Build Structure For Rapid Error Detection and Performance Tuning | ||||
 |
|
|
||
| Graphical depiction of build structure | |
|
|
Instantly spot serializations or inactive nodes in the cluster to identify opportunities for speed improvements |
| Job details | |
|
|
Right click on a job to discover information such as job length, files involved, etc. |
| Show dependencies | |
|
|
Visually highlights dependent jobs |
| ElectricSimulator | |
|
|
Determine how fast a build could run in parallel with additional nodes in the cluster for better resource planning |
| Reports | |
|
|
Unprecedented visibility into the build, including longest serial chain, longest job, serialization analysis, job time by type, etc. |
ElectricInsight — Product Details
ElectricInsight® provides powerful, at-a-glance software build diagnostics. How does it do that?
Traditionally, finding out why a build slowed down, why a build broke, or which files were dependent on others was a time-consuming, manual task. But now, the ElectricInsight tool graphically depicts how the build is structured and run, empowering build managers to pinpoint performance problems or conflicts in a parallel build.
ElectricInsight mines the information produced by Electric Make®, the job distribution engine behind ElectricAccelerator®, to provide an easy-to-understand, graphical representation of the build structure for performance analysis.
It provides detailed information and reports on each job on each node of the build cluster, helping users identify which jobs are being performed, when, and with what files for at-a-glance diagnostics. With a single click, see exactly where dependencies lie. Run a report — are there any unnecessary serializations? Are jobs running locally that could be run on the cluster? With this insight you can get the best possible performance out of your build environment.
Instead of manually pouring over tens of thousands of lines of information in a build output file, error detection and performance tuning that once took hours or days can now be done in seconds. With ElectricInsight you can quickly pinpoint opportunities for better builds.
What's New in ElectricInsight?
ElectricInsight version 3.2 can connect to in-flight ElectricAccelerator builds and display jobs as they execute, showing exactly what commands are being processed on the cluster and letting the user instantly assess build progress. Data visualization is also improved with a new heuristic job classifier, which groups targets by primary tasks and enables better analysis of build performance, and additional reporting features exposing both build content and structure.
New features and enhancements include:
- Runtime build monitor features with real-time as well as after-the-fact analysis options:
- ElectricInsight can monitor a running Electric Make build, allowing you to view its progress and get instantaneous information about what commands each agent is executing.
- Completed builds can be replayed to simulate or analyze execution order.
- New and enhanced reports provide additional insight into software build performance:
- New: Build manifest - lists all files that were read and/or written by the build.
- New: Jobs by agent - lists the number of jobs run by each agent in the build.
- New: Jobs by file - lists which jobs read or wrote a particular file.
- Enhanced: Longest serial chain report - now enables you to display the longest serial chain leading to a specific job.
