Events

Events in Gigantic track user activities and system operations throughout the platform. They provide an audit trail of actions performed within models, projects, and organizations, and serve as the foundation for the notification system in the UI.

Purpose and Usage

Events are used for several key purposes:

Audit Trail: Track all user actions for compliance and historical reference
Notifications: Power the real-time notification system in the user interface
Activity Monitoring: Monitor user and system activity across the platform
Operational Insights: Provide insights into system usage and performance

Event Model Structure

The events are stored in MongoDB with the following structure:

Field	Type	Description
`id`	ObjectId	Unique identifier for the event
`user`	ObjectId (ref: User)	Reference to the user who triggered the event
`project`	ObjectId (optional)	Reference to the project context
`organization`	ObjectId (optional)	Reference to the organization context
`model`	ObjectId (optional)	Reference to the model context
`sourceId`	ObjectId	Identifier of the specific entity that triggered the event
`action`	EventAction	Type of action performed (sign, run, start, end, add, edit, delete)
`source`	EventSource	Origin of the event (Auth, Organization, Project, Model, Rule, Scan, Discover, Dataset, Report, Load, Pump, Entity, Field, Label)
`info`	String	Additional information about the event
`createdAt`	Date	Timestamp when the event was created
`updatedAt`	Date	Timestamp when the event was last updated
`readBy`	[ObjectId]	List of users who have read/mark the event
`meta`	Mixed	Flexible metadata storage for additional event details

Event Actions

Events can have the following actions:

sign - User has signed or approved something
run - A process or job has been executed
start - An operation has started
end - An operation has ended
add - Something has been added
edit - Something has been modified
delete - Something has been removed

Event Sources

Events can originate from the following sources:

Auth - Authentication-related events
Organization - Organization-level events
Project - Project-level events
Model - Model-level events
Rule - Rule execution events
Scan - Scan job events
Discover - Discovery process events
Dataset - Dataset-related events
Report - Report generation events
Load - Data loading events
Pump - Data pumping events
Entity - Entity-related events
Field - Field-related events
Label - Label-related events

Storage and Retention

Events are stored in a MongoDB collection named "events". Based on the code review, there doesn't appear to be an automatic TTL (Time To Live) mechanism configured for events, suggesting they are stored indefinitely until manually removed.

The client-side event store builds navigation links for different event types to direct users to the relevant sections of the application:

Scan/Discover events: Navigate to job details page
Dataset events: Navigate to datasets page (for add action) or job details page
Load/Pump events: Navigate to job details page
Rule events: Navigate to rule details page
Model events: Navigate to model page
Report events: Navigate to audit report page

Real-time Updates

Events are published in real-time using GraphQL subscriptions. The client-side event store subscribes to new events and updates the UI notification system accordingly.