Configuring RCABot and ML Models

Introduction

vuRCA Bot is a comprehensive solution designed to simplify incident detection and root cause analysis. It comprises four distinct layers, each contributing to its functionality.

Overview of vuRCA Bot configuration workflow

vuRCA Bot consists of 4 main layers which are loosely connected to complete the setup.

1. Data Store: In this layer, the user specifies the configuration details and how to get the actual data (time series data/metrics) collected from the various touch points and it is inserted and stored here. This data store could be their internal database or a data lake or cloud in some cases.
   
   

   Since the system needs to access the data in real-time, it requires specific permissions to read and access the data stores. These details typically include host info, user credentials, keys, and tokens if the data is in the cloud. These details will be stored in an encrypted file internally.

   

2. Data Model: A data model is an abstraction of each golden signal with its dimension, transformations, and properties associated with it. Users have to define or configure the data model definition which might also include certain domain-level information such as critical thresholds, and upper & lower bounds of that signal as per business requirement, and create a data model on top of this information.
   
   
3. Workspace: At a high level, the Workplace provides a logical separation between various journey graphs. This way it helps provide vuRCA as per the specified information at the workspace level. Some of the key specifics that can be defined in the workspace are as follows.
   • The role of each signal as leading indicators or user experience signals, operational metrics, influencers, etc.
   • Encapsulation of multiple signals into components and applications
   • Relationship and dependencies among signals. The dependencies capture the strength and directions between them.
     
     
4.  Incidents: This layer is the homepage for all the incidents detected by the vuRCA Bot. Incidents across all active workspaces for various applications at the org level will be displayed here.

Prerequisites:

The user must have configured Data Models and Data Store on vuSmartMaps. Refer vuSmartMaps User Guide to know more about the configuration.

Workspaces

This step is required for reading the information provided from the Data Model and mapping them as part of an application or business journey. 

vuRCA Bot will use this information to detect incidents and provide the probable root cause details for particular outages and incidents.

We have four main features here:

1. RCA Bot: It performs operations on your data and finds the probable root cause of an incident.
   • Details
   • Schema
   • Signalizers
   • Bot Settings
   • Storyboards
2. Offline RCA Bot: It performs operations on your old data set and finds the probable root cause of an incident.
   • Details
   • Schema
   • Signalizers
   • Bot Settings
   • Storyboards
3. Time Series Analysis: It performs operations on your data and detects anomalies, provides insights on the Anomalies, and helps you with Auto Baselining and finding incident clusters.
   • Details
   • Source and Schema
   • Signalizers 
   • Storyboards
4. Event Correlation: It is a sub-module that helps customers optimize their time while investigating potential downtimes and failures inside the application.
   • Basic Details
   • Event Sources
   • Settings and Save
   • Alert Console

Further Reading:

1. RCA Bot
2. Offline RCA Bot
3. Time Series Analysis
4. Event Correlation