How Can We Help?
HEAL is a new-age IT tool that uses AI algorithms and machine learning models to help enterprises run without a hitch. Using a patented technique called ‘workload-behaviour correlation’, HEAL analyses all the aspects that go into the smooth running of an IT system (the cumulative volume, composition, and payload), and reacts every time an abnormal behavior occurs, triggering either a healing action or a scaling action depending on the root cause of the problem.
Basic Entities & Relationships
This section will help you in getting familiar with the Basic entities and relationships in Heal.
Enterprises can contain one or more components which may include Web, Application, and Database. Heal clusters the homogenous components instances and their underlying host as Services (micro-services). Services can make or receive a request from another service to perform various tasks. These request exchanged between the services is termed as Dependencies. The service Dependency Map visualizes the connection and dependency between the services.
You can tag one or more services as an Application and view their health in the Heal User Interface.
For example, let us consider a Travel booking Enterprise named ‘XR Travels’ as in the “SDM-Service Dependency Map” image. XR Travels allows users to book flights and Hotels. Booking is done by Customers, Travel Partners, and Hotel Partners.
Instances of XR Travels’ Web layer are clustered as a ‘Travel Web’ service. Similarly, the instances of the Application layer of Hotels and Flights are clustered to form the Hotels and Flights Services. In the SDM Page, each Circle represents a Service. The connectors represent the request between the services and the arrow denotes the direction in which the request flows.
Refer Viewing Service Dependency Map for more details on SDM.
Time Series Data
Heal captures the Enterprise health in terms of Time series data, termed as Workload KPIs and Behaviour KPIs.
### Workload KPIs
Workload KPIs are the Key Performance Indicators about the amount of load in terms of Request & Response served by the instance, status and response time of the request. Business information captured in the response of Request adds input to the Experience Journey.
Request aka transactions are technical request (For Example: URI call) and received by a Service from an external actor or another service.
With the ACME Travels example,
- Transaction: URI request made to the ‘Travel Web’ by the Customer.
- Request: Request made from one service to another service.
Behaviour KPIs
Behaviour KPIs are Key Performance Indicators about how the monitored instances behave/function. Behaviour KPI can be Performance metric, Availability metric.
Performance Metric KPI: Performance Metric KPI represents the performance information of the component instances, host instances. The values can be numerical values of different types.
Availability KPI: Availability KPIs represent the state of instances, port. The state can be either up or down. The values are binary (0 or 1).
Triggers from MLE
With the help of Machine Learning Engine, HEAL establishes a relationship between the workload and the behaviour of the system. The outcome of the MLE could be an anomaly, early warnings, Incidents and Issues. MLE also triggers Forensic actions and self-healing.
Archived Data Repository
Heal provides you ability to generate customised dashboards, reports suiting the business requirement. Archived Data Repository is available to achieve this.
To summarise, Heal provides self-healing solution to your enterprise by predicting, diagnosing and preventing the problems.