Investigations overview

Prev Next

Start here after Darwinium has processed an interaction. Investigations shows recorded events and their assessment results; Workflows is where you change the configuration that generates them.

Find your first event

  1. Select the node for the environment you tested.
  2. Open Investigations → Query and select a time range covering the interaction.
  3. Search for a known journey or step, such as step_name = "login_submit", using your configured name.
  4. Select Search, then double-click a matching row to inspect its event details.
  5. Check the mapped identifiers, profiling fields and configured signals, scores or decision. Use the journey ID to follow related events.

Use Search events for query controls, Customize the event table for columns, and Follow a journey timeline for session context. If a new integration has produced no events, return to Deployment confirmation.

The Investigations section on Portal is available to investigate and query journeys and events processed by Darwinium.

The investigations functionality is intended for searching and reviewing details about the events processed by Darwinium

Query View

See Query View.

The Query view is the best place to searches and find journeys that have gone through the Darwinium risk decision engine.

Events are searched over using the Darwinium Query Language. 

That is the same language as used in rules, which ensures consistency between:
1. how you can search over events in this Investigations view
2. how you may define a binary rule to run in the Darwinium rule engine

That helps to translate the outcome of an investigation search directly into a real time rule

Q: What is the delay for seeing events in Darwinium after being processed?

It should only take a few seconds after an event is processed by the Darwinium real-time engine to appear in the Investigations Portal

Event Sidebar

The event sidebar can be triggered upon double clicking any row of data in the table.

The event sidebar works best to analyse the specific details about the Darwinium event like:

  • What was the Darwinium decision and why

  • What signals and scores evaluated on the event

  • What was the email/device/ip/location/device forensics on this event

Tip: Full event data

You can also right-click a row of data -> show raw event data to see and search on the full event blob

Event Sidebar: Filtering

Additionally, the event sidebar is a useful way of forming queries without needing to know the search syntax;

Hover over any attribute value in the sidebar to see a menu that adjusts your Investigations Search query automatically.
It will populate with the attribute name and the correct function when searching for that attribute type.

  • Filter:Value (Add to Query): use AND to add this attribute value check to search  

  • Replace Query (Replace Query): delete current search and replace with this attribute value check

  • Filter:Condition (Add to Query): If a derived attribute (like a signal), use AND to add the underlying condition to search

  • Filter:Condition (Replace Query): If a derived attribute (like a signal), delete current serach add replace with the underlying condition to search

Journey View

See Journey View.

Darwinium places emphasis on the continuous session that a user has with your site or mobile application.

The Journey View is perfect for isolating a particular user session from start to finish.

The Data Table and sidebar are present in this view too.

Identifier View

See Identifier View.

The Query and Journey view are useful for the tabular data. But it can help investigations to have a view showing which identifiers and attributes are linked.

Linked meaning, used together at least once on a customer journey processed by Darwinium.

Identifier view can be used to:

  • Search for attribute value and linked identifiers in a period of time

  • Search for attributes by signature similarity

  • Perform a number of hops (attributes values two or three links away from original)

And even perform operations such as:

  • Adding labels to identifier values

  • Export graph data to JSON

Investigate agent and MCP activity

Once MCP and JSON-RPC steps are deployed, use the same investigation tools to inspect tool calls, mapped arguments, response results and errors. Filter to the relevant journey and step; add mapped fields to the event table. A shared journey requires consistent session correlation in the integration.