Overview
Powered by the Device Log Analysis template in Agent Composer—Enterprise; request a demo. Root cause analysis (RCA) for device logs. The agent parses logs, cross-references debug rules, and generates RCA reports—work that would otherwise take a senior engineer hours manually. Demo log files are synthetic (created for this demo); download them from the live demo and try the same workflow in your own environment. The live demo includes three sample datasets—each is a log plus matching debug reference for a different scenario. Run one of the suggested queries (or your own). For more Agent Composer demos beyond DLA, see Contextual AI Demos.Demo behavior: Refreshing the page will clear your progress. For this demo, queries are cached and sped up (production DLA template).
Try the Demo
Launch Demo
Analyze device logs yourself
Sample datasets in the demo
The live demo offers three preloaded options—each is a pair of synthetic files (device log + debug reference). Use the links below for scenario-specific example questions, example outputs, and file descriptions.3GPP Wireless
LTE-style eNB/RRH logs, handovers, CPRI/HARQ/BLER. Example: “Why did the calls drop?”
ATE Board Validation
Multi-site test logs, VR telemetry, DDR5/PCIe. Example: “Why did SITE:2 fail?”
Mazda Infotainment
CMU crash logs, format-string-style bugs. Example: “Why did my car stereo restart?”
The Problem
When a device or network incident occurs, engineers must manually sift through thousands of timestamped log entries containing cryptic error codes, protocol abbreviations, and vendor-specific identifiers. Root cause analysis requires:- Decoding error codes against reference documentation
- Correlating events across subsystems
- Building timelines to establish causality between failures
- Identifying patterns across multiple failure events
- Distinguishing symptoms from root causes
How It Works
In production
What you do: Upload a log file and optionally additional context (e.g. a debug reference guide), then ask a query—for example “Why did the call drop?”. What happens automatically: A multi-agent implementation takes over. The system breaks the work into tasks, shows trajectory as agents run, parses the logs, builds searchable databases from the log and reference files, runs root cause analysis, and produces outputs. Pipeline stages: Uploading files → parsing logs → building the database → root cause analysis → generating the report. Outputs: The agents can generate Python scripts (e.g. custom parsers) that run in a secure VM, produce visuals (causation timelines, degradation charts), and deliver a detailed RCA report (executive summary, timeline, decoded errors, recommended actions). Auditable: You specify the output (report, artifacts, visuals). Every step—tasks, trajectory, intermediate artifacts, and final report—is visible so teams can review and learn from the analysis.In this demo
Three demo datasets are preloaded (3GPP Wireless, ATE Board Validation, Mazda Infotainment). Each is a device log plus matching debug reference. See the dataset pages above for suggested queries and example outputs.Output
Each scenario produces a detailed RCA report and optional visualizations. Report content varies by domain but typically includes an executive summary, fault or session timeline, decoded errors, root cause chain, and prioritized recommended actions. For concrete example outputs and screenshots:- 3GPP Wireless — Call drop causality, system health timelines, BER/BLER/HARQ comparisons
- ATE Board Validation — VDDQ voltage/thermal charts, per-site comparison, tester artifact analysis
- Mazda Infotainment — Incident timeline, crash sequence, format string vulnerability walkthrough
Sample files in the demo
Each of the three datasets is a pair of synthetic files: a timestamped device log and a debug reference guide. The live demo loads all three so you can compare scenarios. For file names and contents per dataset:- 3GPP Wireless —
3gpp_wireless_device_log.txt,3gpp_wireless_debug_rules.txt - ATE Board Validation —
ate_device_log_30mb.txt,ate_debug_rules.txt - Mazda Infotainment —
mazda_device_log_30mb.txt,mazda_debug_rules.txt
Learn More
- Contextual AI Demos — Other live demos (spec search, materials, test automation, and more)
- Getting Started with the Contextual AI Platform
- Example: Device Log Analysis (Enterprise) — How multi-source root cause analysis fits the platform
- Agent Composer Templates — Basic Search, Agentic Search, and Device Log Analysis (Enterprise)
- Request a demo