A non-transitory computer-readable medium having encoded therein programming code executable by one or more processors to perform or control performance of operations of automated diagnosis and corrective action implementation at a managed device based on a device-level evaluation of digital experience metrics, the operations comprising:
- accessing attributes data representative of a plurality of attributes associated with the managed device included in a managed network, the managed network implementing software as a service (SAAS) management operations relative to the managed device including service management operations implemented to provide information technology (IT) support to the managed device, security management operations implemented to provide patch management at the managed device, and application management operations implemented to provide access to particular products at the managed device, wherein:
- each attribute of the plurality of attributes is a digital experience metric used to calculate a digital experience index (DEXI) of a user who is associated with the managed device;
- and
- the attributes data includes a first subset of attribute data derived from operation of the managed device, a second subset of attribute data derived from the security management operations, a third subset of attributes derived from the service management operations, and a fourth subset of attribute data derived from application management operations;
- further accessing operational data directly from the managed device;
- based on the operational data, identifying normal operation of the managed device during which the managed device is in a condition in which components of the managed device function within standard and recommended operational parameter ranges, and anomalous operation of the managed device during which at least one component of the managed device is operating outside the standard and recommended operational parameter ranges;
- correlating the attributes data to the operational data to identify attribute value ranges during the normal operation of the managed device and the attribute value ranges during anomalous operation of the managed device;
- normalizing the attribute value ranges to enable comparisons between the first, the second, and the third subsets of attributes data;
- evaluating interactions between the plurality of attributes relative to one another and relative to the operational data to identify attribute contributions during the normal operation and during the anomalous operation using an unsupervised machine learning model, wherein the attribute contributions are an extent to which each attribute contributes to the anomalous operation or to the normal operation;
- assigning a weight to each attribute of plurality of attributes, wherein the weight reflects the attribute contribution of the attribute to which the weight is assigned during the normal operation;
- determining a normal device index range (NDIR) for the managed device, wherein the NDIR includes a combination of the normalized attribute value ranges during normal operation of the managed device;
- after the NDIR is determined monitoring current attribute data for the plurality of attributes during real time operation of the managed device;
- computing a calculated device index (CDI) as a sum of each of the assigned weights applied to each of the attributes using the current attribute data;
- comparing the CDI to the NDIR;
- and
- responsive to the CDI being outside the NDIR:
- inferring that the managed device is in anomalous operation;
- identifying a first attribute of the plurality of attributes that is outside a first attribute value range of the first attribute;
- further identifying a component of the managed device or of the managed network that affects the first attribute;
- and
- automatically implementing a corrective action on the component to modify the component such that the first attribute is within the first attribute value range and the CDI is within the NDIR.