Adding Autodiscovery to a KPI Analysis¶
Autodiscovery provides the opportunity to set up in advance, a KPI analysis that can run against an advanced KPI analysis configured with a threshold; when the threshold is crossed by the analysis data, the autodiscovery automatically runs and saves its results.
Manual Troubleshooting vs. Using Autodiscovery¶
Normally, when you receive a notice that a KPI analysis threshold has been crossed, your first response would be to troubleshoot the issue. To troubleshoot, you would create a new KPI discovery analysis with the same parameters as the analysis whose threshold was crossed. Thus, by manually creating and running a new discovery analysis against the results of the first analysis, you jumpstart the process of determining which fields and factors caused the data to cross the threshold.
An autodiscovery analysis is created using the same steps as used in creating any advanced KPI. See How to Create an Advanced KPI, here.
The autodiscovery feature allows you to set up this troubleshooting analysis ahead of time, and link it to the first analysis that you want to troubleshoot. With the two analyses linked, the system now triggers the troubleshooting analysis to run automatically when data in the first analysis crosses the threshold you set.
This feature allows you to set up a relationship between two analyses; in the first analysis, you set a threshold so the system notifies you if the analysis results cross it; you set the second analysis up to run against the results of the first analysis when results cross the threshold. You set the second analysis up as a troubleshooting analysis that runs automatically, in essence, to assist you in determining causal factor(s) of the first analysis exceeding the threshold.
Example: For explanatory purposes, let's call the two analyses 'DiscoveryAnalysis 1' (DA1) and 'Discovery Analysis 2' (DA2). There is nothing special or different about these two discovery analyses, other than setting a threshold in DA1. If the threshold is crossed when DA1 runs, you receive a notification about it. This also triggers DA2 (the troubleshooting analysis) to run against the results of DA1.
Helpful Naming Convention¶
Following this naming convention when you create and save an analysis configured with autodiscovery can help you identify it on the KPI list page.
If you name the first analysis in this relationship, "Line3Yield", it is helpful to give the second analysis (which is used to troubleshoot the first analysis) a name that is similar to the first analysis, such as "Line3YieldTRBLshoot", or something similar for two reasons, so:
- the two related analyses appear adjacent to one another in the KPI list
- you can easily distinguish troubleshooting analyses from other analyses
How to Configure the Autodiscovery Settings¶
This section presumes that you have already configured the schedule, share, and notify settings for your analysis.
Follow these steps to add autodiscovery to your analysis:
- To automatically run an additional discovery analysis when data in this KPI analysis crosses the threshold, select a KPI discovery analysis from the "Autodiscovery" drop-down list.
- Click "Save". The system saves your KPI and displays a Success message.