Observability Studio

Observability Studio is a multi-accelerator platform for observability data extraction, maturity assessment, and operations analysis.

It combines:

• A FastAPI backend for extraction and report generation
• A React Hub UI for module-driven workflows
• Reusable adapter-based extraction through the ObservaScore Common Observability Model (COM)

What This Repo Is Today

This repository has evolved from a single CLI assessment tool into a platform with five active modules in the Hub UI:

1. ObsCrawl: Crawl and export observability estate data to Excel
2. ObservaScore: Deterministic maturity scoring with optional AI narrative
3. RCA Agent: Incident investigation and blast-radius analysis
4. RED Panel Intelligence: Service-centric RED dashboard coverage quality
5. Observability Gap Map: Application/service coverage mapping with debugging-path connectivity checks

Core Capabilities

• Read-only data extraction from multiple observability tools
• Deterministic scoring for coverage/readiness workflows
• Optional AI enrichment for narrative sections (non-required)
• Offline-shareable HTML reports plus JSON artifacts
• Feature-flag based accelerator enable/disable at runtime

Supported Tool Adapters

Current adapter set includes:

• Prometheus
• Grafana
• Loki
• Jaeger
• Alertmanager
• Tempo
• Elasticsearch
• Datadog
• Dynatrace
• AppDynamics
• Splunk
• OTel Collector

Platform Architecture

Request flow:

React Hub UI -> FastAPI routes -> Service layer orchestration -> ObservaScore extraction (COM) -> Accelerator analysis logic -> Runtime artifact generation (HTML/JSON/XLSX) -> Download/preview endpoints

Backend layers:

• Routes: request validation and HTTP contract
• Services: orchestration and runtime folder handling
• Accelerators: deterministic analysis logic and report generation
• Adapters: tool-specific read-only extraction

API Surface

Health and Platform:

• GET /api/health
• GET /api/feature-flags

Hub v1 endpoints:

• POST /api/v1/validate
• POST /api/v1/crawl
• POST /api/v1/assess
• POST /api/v1/rca

Current platform endpoints:

• POST /api/observability-gap-map
• POST /api/red-intelligence

Legacy compatibility endpoints:

• POST /api/export
• POST /api/assess

Artifacts:

• GET /api/download/runtime/{path}
• GET /api/preview/runtime/{path}

Observability Gap Map: Current Behavior

Gap Map remains focused on service-level signal coverage and now includes a separate debugging-path connectivity layer.

Coverage layer (existing):

• Metrics, logs, traces, dashboards, alerts, RED readiness per service
• Coverage scoring, readiness bands, missing signal recommendations

Signal Connectivity layer (new):

• Separate section in analysis/report, not merged into coverage matrix
• Service-level connectivity checks:
  • metrics_to_logs
  • logs_to_traces
  • alerts_to_dashboards
  • dashboards_to_logs
  • dashboards_to_traces
• Deterministic scoring:
  • PASS = 100
  • WARN = 60
  • FAIL = 0
• MTTR risk classification:
  • low: score >= 80
  • medium: score >= 50 and < 80
  • high: score < 50
• Added JSON fields:
  • connectivity_results
  • connectivity_summary

Frontend Hub

Main Hub tile inventory:

• obscrawl
• observascore
• rca_agent
• red_panel_intelligence
• observability_gap_map

Feature flags control tile and API availability through:

• platform/config/feature_flags.yaml
• backend middleware enforcement in backend/app/main.py

Running Locally

Prerequisites:

• Python 3.10+
• Node.js 18+

Install backend dependencies:

pip install -e .
pip install -r backend/requirements.txt

Run backend:

uvicorn backend.app.main:app --host 0.0.0.0 --port 8001 --reload

Run frontend:

cd ui
npm install
npm run dev

Configuration

Runtime extraction config is generated per execution under runtime/<run_id>/.

Tool connection details are submitted from UI payloads and converted into runtime config files by backend services.

Splunk-specific URL derivation and auth mapping are supported in the config builder/service payload path.

Outputs

Generated artifacts are written to runtime/<run_id>// and typically include:

• HTML report (primary preview/download target)
• JSON report (structured output)
• XLSX export (for crawl/export workflows)

Repository Map (High Level)

• accelerators/: domain logic for obscrawl, observascore insights, rca-agent
• backend/app/: FastAPI app, routes, services, schemas
• ui/src/: Hub UI, modals, API client, styles
• shared_core/: shared flags and platform internals
• platform/config/: feature flag configuration
• runtime/: generated outputs (git-ignored in normal workflows)

Notes

• All extraction is read-only toward source tools.
• Deterministic analysis is the baseline; AI is additive and optional where enabled.
• Existing legacy APIs are retained for compatibility while Hub v1 endpoints handle streamlined module flows.