How shashivardhan-7/deploydetective-exasol works
Deploy Detective is a web-based AIOps agent that investigates microservice incidents using Exasol, recommends a safe response, and verifies recovery. The principal workflow starts with a user request, resolves the affected service, analyzes error rates, correlates deployments, checks dependency latency and historical incidents, then decides whether to simulate a rollback and query for recovery. The README documents logs, metrics, deployment history and incident history as analytical inputs. The graph follows the sampled application and agent code; SQL capabilities and database structure are supported by the README and tree, but their contents were not sampled.
Main entry points: tools.py Exasol query and action tools · investigator.py Investigation orchestrator · main.py FastAPI application · schema.sql Exasol analytical store
Components 16
Web interface
- Investigation UI
static/app.js - FastAPI application
app/main.py
Investigation agent
- Investigation orchestrator
agent/investigator.py - Service and intent planner
agent/planner.py - Remediation policy
agent/decision.py - Evidence narrative
agent/narrative.py
Exasol analytics
- Exasol query and action tools
agent/tools.py - Analytics query definitions
database/queries.sql - Error-rate analysis
database/queries.sql - Deploy correlation
database/queries.sql - Dependency latency
database/queries.sql - Historical incident match
database/queries.sql
Action and verification
- Simulated rollback adapter
deployment/mock_deployer.py - Recovery verification
database/queries.sql
Other components
- Incident requester
- Exasol analytical store
database/schema.sql
Connections 24
- Incident requester → Investigation UI: submits incident
- Investigation UI → FastAPI application: posts request
- FastAPI application → Investigation orchestrator: runs investigation
- FastAPI application → Exasol query and action tools: checks health
- Investigation orchestrator → Service and intent planner: resolves service
- Investigation orchestrator → Exasol query and action tools: requests analysis
- Exasol query and action tools → Analytics query definitions: loads queries
- Exasol query and action tools → Error-rate analysis: runs query
- Exasol query and action tools → Deploy correlation: runs query
- Exasol query and action tools → Dependency latency: runs query
- Exasol query and action tools → Historical incident match: runs query
- Exasol query and action tools → Exasol analytical store: executes SQL
- Exasol analytical store → Investigation orchestrator: returns evidence
- Investigation orchestrator → Remediation policy: evaluates evidence
- Investigation orchestrator → Evidence narrative: builds explanation
- Investigation orchestrator → Exasol query and action tools: records rollback
- Exasol query and action tools → Simulated rollback adapter: proposes rollback
- Exasol query and action tools → Exasol analytical store: writes action audit
- Investigation orchestrator → Exasol query and action tools: queries recovery
- Exasol query and action tools → Recovery verification: runs query
- Exasol query and action tools → Exasol analytical store: reads telemetry
- Exasol analytical store → Investigation orchestrator: returns verification
- Investigation orchestrator → FastAPI application: returns result
- FastAPI application → Investigation UI: returns investigation