AppDynamics switched from Semantic Versioning to Calendar Versioning starting in February 2020 for some agents and March 2020 for the entire product suite.

    Skip to end of metadata
    Go to start of metadata

    The AppDynamics KPI analyzer helps you to identify the root causes of poor application performance. By applying machine learning algorithms, the KPI analyzer recognizes performance anomalies based on the average response time key performance indicator (KPI) for a node. It isolates the metrics that are the most likely contributors to the poor performance and identifies the likely degree of impact on the KPI for each metric.

    In this way, the analyzer makes troubleshooting quicker and easier by surfacing those metrics that are most likely to be related to the root cause for poor response times.   

    The KPI analyzer correlates metrics to performance issues using metric data from the observed node from the last 4 hours. It ranks the contributing metrics based upon likely impact using a contribution score. The contribution score indicates the degree of impact of the underlying metric on the KPI.  

    Enabling the KPI Analyzer

    The KPI Analyzer is available on select SaaS Controllers only. If you are interested in using the feature, contact your AppDynamics representative. 

    Using the KPI Analyzer

    Users can view KPI Analyzer information by drilling down into node information for a transaction snapshot.

    The general workflow you follow to perform root cause analysis with the KPI Analyzer is as follows: 

    1. From a transaction snapshot list, find the transaction snapshot for the transaction to view: 

    2. In the snapshot viewer, drill down into the specific node that is presenting performance problems: 

    3. Click the KPI Analyzer tab. 

      The tab shows the KPI and contributing metrics as time series charts, as follows:

      The tab is organized into two general areas: 
      • The top of the tab shows the key KPI, average response time, values over time. 
      • The metrics that are identified as most likely contributing to the performance are shown in ranked order below, allowing you to identify and resolve the root causes of the performance problems. 
    • No labels