How bhaskar1001/agentic-ai-incident-platform works
This is a Python incident-response agent: a caller submits an incident description, the workflow assesses severity, investigates the incident, then chooses escalation or monitoring. Investigation uses a bounded model/tool loop to gather mock logs, metrics, and pod status, preserving evidence and a timeline in structured results. The incident domain also defines status-transition rules, though the sampled workflow does not use the Incident entity. The README is empty; unsampled API details and persistence are not represented.
Main entry points: incident.py Incident workflow · investigation.py Investigation agent · severity.py Severity assessment · registry.py Tool registry
Components 11
Incident workflow
- Incident workflow
src/agentic_ai/workflows/incident.py - Monitoring outcome
src/agentic_ai/workflows/incident.py
Investigation
- Investigation agent
src/agentic_ai/agents/investigation.py
Domain models
- Incident lifecycle
src/agentic_ai/domain/incident.py - Investigation results
src/agentic_ai/domain/investigation.py
Observability tools
- Tool registry
src/agentic_ai/tools/registry.py - Mock observability tools
src/agentic_ai/tools/mock_tools.py
Model integration
- Severity assessment
src/agentic_ai/llm/severity.py
Other components
- Incident caller
- Ollama model service
- Human responder
Connections 11
- Incident caller → Incident workflow: submits incident
src/agentic_ai/workflows/incident.py - Incident workflow → Severity assessment: assesses severity
src/agentic_ai/workflows/incident.py - Severity assessment → Ollama model service: requests assessment
src/agentic_ai/llm/severity.py - Severity assessment → Incident lifecycle: uses severity type
src/agentic_ai/llm/severity.py - Incident workflow → Investigation agent: runs investigation
src/agentic_ai/workflows/incident.py - Investigation agent → Ollama model service: reasons and summarizes
src/agentic_ai/agents/investigation.py - Investigation agent → Tool registry: dispatches tools
src/agentic_ai/agents/investigation.py - Tool registry → Mock observability tools: calls registered tools
src/agentic_ai/tools/registry.py - Investigation agent → Investigation results: records evidence and results
src/agentic_ai/agents/investigation.py - Incident workflow → Human responder: escalates urgent cases
src/agentic_ai/workflows/incident.py - Incident workflow → Monitoring outcome: routes normal cases
src/agentic_ai/workflows/incident.py