Skip to main content

​
Overview

This agent performs automated root cause analysis on 3GPP LTE device failure logs. It cross-references raw device logs against a debug reference guide to decode error codes, build event timelines, and identify causal chains—work that would normally require a senior telecom engineer hours of manual investigation. Powered by the Device Log Analysis template in Agent Composer—available for Enterprise customers; request a demo for access.

​
Try the Demo

Launch Demo

Analyze device logs yourself
 3GPP samples preloaded—preview the files, pick a suggested query, and run the analysis.
Running on the production DLA template. Demo queries are cached and sped up for demonstration purposes.

​
The Problem

When an LTE 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 (EMM cause codes, X2AP causes, S1AP causes) against reference documentation
  • Correlating events across subsystems (RRC, NAS, PHY, MAC, X2AP, S1AP)
  • 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

Sample 3GPP files are preloaded; choose a suggested query and run the analysis.

​
Example Questions

Suggested queries in the demo:
  • “Why did the call drop?”
  • “Which handovers failed and why?”
  • “What is the significance of the thermal readings in this log?”

​
Output

Example outputs: Call drop causality timeline — Swimlanes (Infrastructure, Core Network, Handover, Call Drops) with causal chains from events (e.g. CPRI degradation, ANR neighbor stale, HO preparation failure) to T304 expiry / RLF (call drops). Shows how the agent traces cause-effect across network layers: Call Drop Causality Timeline — swimlanes and causal chains to call drops System health timeline — Multiple correlated metrics over time (CPRI fronthaul BER, baseband temperature and throttle, PRB utilization, uplink BLER, S1-U backhaul jitter) with thresholds. Illustrates the range of metrics the agent can surface and correlate: System Health Timeline — BER, temperature, PRB, BLER, and jitter over time The final RCA report (e.g. for “Why did the call drop?”) includes:
  • Executive summary — Root cause (e.g. CPRI fronthaul degradation on RRH:2), secondary factors (DRX misalignment, SCTP/MME, backhaul jitter), and impact (UE counts by failure type: RLF, RRC reject, NAS attach reject)
  • Detailed timeline — Phased chronology with decoded log lines (early degradation → threshold breach → handover failures → thermal throttle → congestion and rejects), so reviewers can follow the sequence event-by-event
  • Error analysis by UE — Grouped tables (e.g. handover failure → RLF call drops, RRC reject due to congestion, NAS attach reject with cause codes) with timestamps and preceding causes
  • Technical explanation — Root cause deep dive and cascade effects (e.g. CPRI BER progression, RF/UL BLER, thermal throttle, ANR staleness and HO prep failure, SCTP/MME overload) with supporting tables
  • Answer to original question — Direct answer to the query (e.g. call drop mechanism: CPRI → RF degradation → thermal throttle → HO failures → T304 expiry → RLF) and optional mechanism diagram
  • Attached visualizations — References to the generated timeline visuals (system health, call drop causality)
  • Recommended actions — Prioritized table (e.g. CRITICAL / HIGH / MEDIUM) with action and rationale for each

​
3GPP sample files in the demo

In the demo, two 3GPP LTE sample files are preloaded; you can preview the files, then run the analysis.
  • Device log (3gpp_wireless_device_log.txt) — A timestamped eNB/UE failure log with entries from subsystems such as eNB-SYS, eNB-MAC, eNB-PHY, eNB-RRC, eNB-S1AP, eNB-X2AP, and UE-MAC/UE-RRC. Each line includes severity (INFO/WARN/ERROR), tags, and codes (e.g. RACH_INIT, RRC_CONN_REQ, CPRI_LINK_DEGRADED_0xE1, VNDR_THERMAL_0x4F2A) that need to be decoded against a reference.
  • Debug reference guide (3gpp_wireless_debug_rules.txt) — An eNB debug reference (e.g. “3GPP LTE eNB Debug Reference Guide”) that defines RRC/NAS procedures, EMM/ESM cause codes, S1AP/X2AP codes, timers, severity levels, RSRP/RSRQ thresholds, and vendor-specific diagnostics. The agent uses this to decode every code and abbreviation in the device log.

​
Learn More