add-it-ya/drdoom

How add-it-ya/drdoom works

DrDoom is an autonomous incident-response service: submitted metric windows are scored, classified, investigated against operational documentation, and turned into a risk-rated remediation held at a human approval gate before execution; reporting and audit preserve the outcome. The API starts and resumes the investigation graph, which owns stage ordering and routing. The repository also includes telemetry preparation and detector/classifier training, plus retrieval evaluation. Relationships beyond the excerpts are included only where documented in the README or named by the SOURCE INDEX; unsampled internals are not assumed.

Main entry points: graph.py Investigation graph · diagnosis.py Diagnosis agent · remediation.py Remediation agent · factory.py Service assembly

Components 23

Telemetry and features

Synthetic telemetry src/drdoom/data/synthetic.py
SMD telemetry src/drdoom/data/smd.py
Metric schema src/drdoom/data/schema.py
Windowing src/drdoom/data/windows.py

Detection and classification

Anomaly detectors src/drdoom/detect/conv_autoencoder.py
Cause classifier src/drdoom/classify/features.py

Investigation workflow

Investigation graph src/drdoom/agents/graph.py
Triage and classification src/drdoom/agents/triage.py
Diagnosis agent src/drdoom/agents/diagnosis.py
Remediation agent src/drdoom/agents/remediation.py
Risk assessor src/drdoom/agents/risk.py
Postmortem writer src/drdoom/agents/reporting.py

Documentation and models

Operations corpus src/drdoom/rag/corpus.py
Document chunking src/drdoom/rag/ingest.py
Hybrid retrieval src/drdoom/rag/index.py
Passage reranker src/drdoom/rag/rerank.py
LLM providers src/drdoom/llm/factory.py

Service and governance

Window API src/drdoom/api/main.py
Service assembly src/drdoom/api/factory.py
Approval and execution src/drdoom/executor.py
Append-only audit src/drdoom/audit.py
Observability src/drdoom/observability.py

Other components

On-call engineer

Connections 28

On-call engineer → Window API: submits windows README.md
Window API → Investigation graph: starts or resumes src/drdoom/api/main.py
Service assembly → Synthetic telemetry: generates traffic src/drdoom/api/factory.py
Service assembly → Anomaly detectors: builds detector src/drdoom/api/factory.py
Service assembly → Investigation graph: assembles service src/drdoom/api/factory.py
SMD telemetry → Metric schema: loads as src/drdoom/data/smd.py
Synthetic telemetry → Metric schema: generates as src/drdoom/data/synthetic.py
Windowing → Metric schema: windows series src/drdoom/data/windows.py
Anomaly detectors → Windowing: scores windows README.md
Triage and classification → Anomaly detectors: checks anomaly src/drdoom/agents/triage.py
Triage and classification → Cause classifier: classifies cause src/drdoom/agents/triage.py
Investigation graph → Triage and classification: runs triage src/drdoom/agents/graph.py
Investigation graph → Diagnosis agent: runs diagnosis src/drdoom/agents/graph.py
Investigation graph → Remediation agent: requests plan src/drdoom/agents/graph.py
Investigation graph → Risk assessor: assesses risk src/drdoom/agents/graph.py
Investigation graph → Approval and execution: gates execution src/drdoom/agents/graph.py
Investigation graph → Postmortem writer: writes report src/drdoom/agents/graph.py
Operations corpus → Document chunking: provides documents src/drdoom/rag/ingest.py
Diagnosis agent → Hybrid retrieval: searches evidence src/drdoom/agents/diagnosis.py
Diagnosis agent → Passage reranker: reranks shortlist src/drdoom/agents/diagnosis.py
Diagnosis agent → LLM providers: generates explanation src/drdoom/agents/diagnosis.py
Remediation agent → Hybrid retrieval: searches guidance src/drdoom/agents/remediation.py
Remediation agent → Passage reranker: reranks shortlist src/drdoom/agents/remediation.py
Remediation agent → LLM providers: proposes fix src/drdoom/agents/remediation.py
Risk assessor → LLM providers: assesses plan src/drdoom/agents/risk.py
Approval and execution → Append-only audit: records decision README.md
Window API → Observability: reports health src/drdoom/api/main.py
On-call engineer → Approval and execution: approves plan README.md

Drawn by GitDiagram on October 1, 2026 from the repository's file tree, README and source files. Read as Markdown ·