LogX

What you will do: understand what LogX adds on top of collected logs and how to get its first report.

LogX applies machine learning and AI to logs to surface what a person would not find by searching: new kinds of errors, changes in the shape of a stream, patterns that repeat before an incident.

What LogX produces

• Anomaly detection — flags events and metrics that deviate from the learned normal for a source.
• Clustering — groups similar log lines into clusters, so a million lines become a few hundred patterns with counts.
• Patterns — tracks when a pattern appears for the first time, disappears or changes rate.
• Daily detected-error reports — a daily summary of the errors detected across sources, new ones first.
• Data stream anomaly alerts — alerts when a stream's volume, rate or structure changes unexpectedly, for example a source going quiet or a new field appearing.

How it fits

LogX analyses the logs collected into your deployment. It is licensed per organization and deployment; AI-driven analysis consumes XMC AI tokens, which an evaluation license includes. See How licensing works.

Steps: first report

1. Enable LogX on a deployment that already collects data:

Console path: <placeholder>

2. Select the sources to analyse. Start with the sources where incidents usually appear.
3. Allow a learning period so the models see normal days:

Learning period: <placeholder>

4. Open the daily detected-error report when it arrives, and the clusters view for a source you know well. Confirm that the clusters match how you would group the lines yourself.
5. Enable data stream anomaly alerts for the sources that must never go quiet.

Reading the results

Treat anomalies as questions, not verdicts. Each one links to the events behind it so you can confirm in search. Clusters and patterns are most useful when you compare today with the last normal day.

Next: Audity

Verify with XPLG engineering before publishing.