ashmitsharma972/project_for_fun

How ashmitsharma972/project_for_fun works

IncidentMind is an incident-response assistant: alerts enter through FastAPI, are classified and persisted, then engineers can request evidence-grounded root-cause recommendations and submit verified feedback. Recommendations are advisory; approval only records simulated approval, not recovery execution. SQLite stores incident and feedback records, while the official Hindsight service is the intended persistent memory layer and Groq is optional. The README also describes dashboard, learning, reflection, replay, anti-runbook, and risk/chaos capabilities. Those latter modules are represented from the tree and documentation; their internal calls were not sampled, so uncertain links are omitted.

Main entry points: api.py FastAPI routes · hindsight_client.py Hindsight adapter · root_cause_analyzer.py Root-cause analyzer · incident_detector.py Incident detector

Components 23

API and ingestion

FastAPI routes backend/api.py
Incident detector incident_processing/incident_detector.py
Incident classifier incident_processing/incident_classifier.py
Log analyzer incident_processing/log_analyzer.py
Incident database backend/database.py

Incident analysis

Root-cause analyzer agent/root_cause_analyzer.py
Runbook recommender agent/runbook_recommender.py
Groq language model agent/groq_llm.py
Runbook content runbooks

Memory and learning

Hindsight adapter hindsight_memory/hindsight_client.py
Incident and evidence models hindsight_memory/models.py
Hindsight service

Learning and experiments

Feedback processor learning/feedback_processor.py
Outcome statistics learning/outcome_stats.py
Reflect insights innovation/reflect_insights.py
Replay Lab innovation/replay_lab.py
Anti-runbook warnings innovation/anti_runbooks.py
Staleness and conflicts innovation/staleness.py
Deploy risk innovation/deploy_risk.py
Chaos Dojo innovation/chaos_dojo.py

Engineer interface

Streamlit dashboard frontend/dashboard.py

Other components

Engineer
Alert source

Connections 25

Alert source → FastAPI routes: submit alerts README.md
Engineer → FastAPI routes: request analysis backend/api.py
FastAPI routes → Incident detector: detect incident backend/api.py
Incident detector → Incident classifier: classify alert incident_processing/incident_detector.py
Incident detector → Log analyzer: analyze logs incident_processing/incident_detector.py
FastAPI routes → Incident database: persist and read backend/api.py
FastAPI routes → Root-cause analyzer: request analysis backend/api.py
Root-cause analyzer → Hindsight adapter: recall evidence agent/root_cause_analyzer.py
Root-cause analyzer → Runbook recommender: recommend steps agent/root_cause_analyzer.py
Root-cause analyzer → Groq language model: request completion backend/api.py
Runbook recommender → Runbook content: select guidance
Hindsight adapter → Hindsight service: retain, recall, reflect hindsight_memory/hindsight_client.py
Hindsight adapter → Incident and evidence models: translate records hindsight_memory/hindsight_client.py
FastAPI routes → Feedback processor: process feedback backend/api.py
Feedback processor → Incident database: store feedback learning/feedback_processor.py
Feedback processor → Hindsight adapter: retain verified outcome README.md
FastAPI routes → Outcome statistics: fetch statistics backend/api.py
FastAPI routes → Streamlit dashboard: serve API data README.md
Engineer → Streamlit dashboard: view evidence README.md
Reflect insights → Hindsight adapter: use reflection innovation/reflect_insights.py
Replay Lab → Hindsight adapter: compare memory modes README.md
Anti-runbook warnings → Hindsight adapter: use retained outcomes README.md
Staleness and conflicts → Hindsight adapter: check memory evidence
Deploy risk → Hindsight adapter: consult incident memory innovation/deploy_risk.py
Chaos Dojo → FastAPI routes: exercise incident flow

Drawn by GitDiagram on November 13, 2026 from the repository's file tree, README and source files. Read as Markdown ·