Optimizing Faulted Job Diagnostics with Intelligent AI Agent Automation
Diagnosing failed automation jobs in UiPath was a highly manual process involving time-consuming log retrieval, parsing of unstructured data, and interpretation of technical messages. IT and CoE teams often struggled with inconsistent log formats, complex JSON structures, and limited real-time visibility — delaying root cause analysis and resolution. These delays disrupted automation reliability and created support bottlenecks across environments.
To eliminate these challenges, Accelirate deployed a fault-tolerant
AI Agent that autonomously retrieves, parses, and summarizes UiPath faulted job logs, empowering teams to resolve issues faster with structured insights and zero manual effort.
01 - Dynamic Job Log Filtering & Retrieval
The AI Agent connects to UiPath Orchestrator using secure API calls, applying filters such as JobStatus and ProcessName to fetch the first faulted job for analysis—no manual lookups required.
02 - Real-Time JSON Log Parsing
Robot logs are retrieved in JSON format and automatically parsed into structured tables using Newtonsoft.Json, transforming complex data into readable outputs within seconds.
03 - Chronological Log Sorting
Log entries are sorted by timestamp ticks to ensure a clear, step-by-step sequence of events, aiding quick understanding and pinpointing of fault origins.
04 - Human-Readable Output Generation
Parsed logs are formatted into a concise, readable string (out_RobotLogs) for easy interpretation, which can be sent via Teams, Email, or integrated dashboards.
05 - Plug-and-Play Agent Design
Built as a reusable backend for larger diagnostic or support bots, the agent accepts flexible input parameters and returns standardized output for consistent use.
06 - Seamless Integration with Monitoring Tools
Can be triggered through Maestro, Action Center, or other monitoring platforms to provide just-in-time diagnostics when a fault is detected in production environments.
07 - Secure & Compliant Data Handling
All data is processed in-memory over HTTPS using authenticated tokens, ensuring role-based access and compliance with enterprise RPA security protocols.
Architecture Overview:
Layer |
Description |
Input Layer |
Inputs received via Agent interface or Maestro (e.g., {{JobStatus}}, {{ProcessName}}, optional date/time filters). Triggered manually or automatically via fault-detection rules. |
Execution Logic (Agent Core) |
The agent invokes the Get Robot Logs From Jobs workflow deployed in Orchestrator. This workflow constructs dynamic OData queries based on inputs. |
API Integration |
Uses HTTP Request activities to interact with Orchestrator’s /odata/Jobs and /odata/RobotLogs endpoints. |
Data Parsing & Transformation |
Robot logs are returned as JSON, then parsed using Newtonsoft.Json. Data is stored in a DataTable with all log attributes (e.g., Level, Message, Timestamp). |
Analysis & Output Layer |
The log entries are sorted by timestamp, formatted, and returned as a text string ({{out_RobotLogs}}). This output can be consumed by Maestro, other agents, or humans. |
Optional Output Destinations |
Results may be shown in Maestro dashboards, UiPath Apps, Action Center, or routed to Slack, email, or MS Teams via downstream agents. |
What Was Unique About the Solution Implemented?
Accelirate’s AI-powered agent for faulted job analysis delivered a highly specialized automation tailored for UiPath Orchestrator environments. Unlike generic log readers, this solution dynamically filtered jobs by status and process name, retrieved the first relevant execution, and parsed complex JSON logs into structured, human-readable summaries — all without requiring developer intervention. The
agent’s architecture featured resilient
API integrations, automated JSON deserialization, and timestamp-based sorting to preserve execution sequence. Designed with reusability in mind, it can plug into broader support workflows, self-healing automations, or triage dashboards. With built-in flexibility for schema modifications and log filtering enhancements, the solution is ready to scale across environments and evolving diagnostic needs.