Log Management - AI Anomaly Logs

The AI Anomaly Logs feature in Ceburu helps users identify, analyze, and investigate unusual log patterns detected by the AI model. Any log entry identified as anomalous is surfaced in this along with contextual insights, root cause analysis, and configuration controls. 

Purpose of AI Anomaly Logs

• Automatically detect abnormal or unusual log behavior

• Visualize anomaly trends over time

• Drill down into individual anomalous log events

• Provide AI-generated root cause analysis

• Allow users to configure model behavior using keywords

Navigate to: Log Management - AI Anomaly Logs

This page displays:

• A time-series graph of log activity

• A detailed anomaly table

• An anomaly details panel

• Model configuration options

Log Activity Graph

The graph at the top provides a visual comparison of:

• Total Logs 

• Anomaly Count (red line)

Each data point represents a specific timestamp.

What this shows:

• Spikes in total logs

• Corresponding increases in anomalies

• Time periods with unusual behavior

Hovering over a data point displays:

• Timestamp

• Total log count

• Number of detected anomalies

You can filter anomalies using Identifier tags, such as:

• Folder

• Source

• Custom identifiers

This helps isolate anomalies related to specific services, folders, or sources.

Anomalies Table

The Anomalies table lists all detected anomalous log entries.

Table Columns

• Timestamp - When the anomaly occurred

• Summary - A preview of the log message

View Anomaly Details

Clicking the View icon on any anomaly opens the Anomaly Details panel on the right.

This panel provides three tabs:

1. Root Cause Analysis 

The Root Cause Analysis tab explains why the log was classified as anomalous.

What it includes:

• AI-generated explanation of the anomaly

• Possible causes (e.g., missing fields, logging changes, abnormal patterns)

• Contextual interpretation of the log deviation

Remediation Steps

The system also provides recommended remediation actions, such as:

• Reviewing logging configuration

• Verifying required fields

• Checking recent code or framework changes

• Monitoring logging behavior in real time

2. Document JSON

The Document JSON tab displays the raw log document in JSON format.

Features:

• Full structured log payload

• Searchable fields and values

• Copy JSON option for external analysis

• Useful for debugging, exports, and integrations

3. Document Table