Application Resilience Testing Tool - User Guide

A browser-based tool for building operational resilience through workflow-driven health modeling. It walks you through four phases - from Failure Mode Analysis to Azure deployment and continuous governance.

📺 Watch the video walkthrough

Contents

_AppResilienceTestingTool/
├── fma-framework.html      ← Main application (open in browser)
├── seq-import.js           ← Sequence-diagram import engine
├── block-import.js         ← Data-flow / architecture-image import engine
├── Get-FoundryToken.ps1    ← PowerShell helper for Azure AI bearer tokens
└── flows/
    ├── 01_ClientOnboarding_seqdiag.mmd   ← Sample Mermaid sequence diagram
    ├── 02_E2E_ClientInt_seqdiag.mmd      ← Sample Mermaid sequence diagram
    └── Overall-Data-Flow.png             ← Sample architecture / data-flow image

Prerequisites

Requirement	Details
Browser	Modern Chromium-based browser (Edge, Chrome) recommended
Internet	Required on first load for CDN libraries (SheetJS, Tesseract.js, Dagre, Google Fonts)
Azure subscription	Needed only for Phase 3 Deploy and AI Vision features
PowerShell 5.1+	Needed only if using Get-FoundryToken.ps1 for AI vision tokens

Quick Start

1. Open fma-framework.html in your browser.
2. The app loads with an empty workspace - no server needed.

---

Step-by-Step Walkthrough

1 - Import Sequence Diagrams (MMD Files)

Use this to import Mermaid sequence diagrams that define your application workflows.

1. Click 📋 Import Sequence Diagram in the toolbar.
2. In the import dialog, choose Upload .mmd file(s).
3. Select the sample files from the flows/ folder:
   • 01_ClientOnboarding_seqdiag.mmd
   • 02_E2E_ClientInt_seqdiag.mmd
4. The engine parses the diagrams, detects Azure services, classifies workflows, and auto-populates Phase 1 data (components, failure modes, RPV scores).
5. Review the imported workflows in the Phase 1 panel.

2 - Import Data Flows (Architecture Image)

Use this to extract workflows from an architecture / data-flow diagram image via AI vision.

1. Click 📊 Import Data Flows in the toolbar.
2. Choose Upload Image and select flows/Overall-Data-Flow.png.
3. The tool sends the image to an AI vision model which extracts sequence diagrams from the architecture.
4. Preview the extracted diagrams in the grid, then confirm to import.

AI Vision requires a bearer token. If prompted, use the helper script:

.\Get-FoundryToken.ps1 -ClientId "<app-id>" -ClientSecret "<secret>" -TenantId "<tenant>"

The token is copied to your clipboard - paste it into the app's AI Vision Settings → Bearer Token field.

Alternatively, create a .env file alongside the script:

FOUNDRY_CLIENT_ID=<app-id>
FOUNDRY_CLIENT_SECRET=<secret>
FOUNDRY_TENANT_ID=<tenant>

3 - Phase 1: Failure Mode Analysis

After import, Phase 1 is pre-populated with:

• Workflows - classified as User Flow or System Flow
• Components - categorised as Application or Platform, with Azure service detection
• Failure Modes - per-component, each scored with RPV (Risk Priority Value = Impact × Likelihood × Detectability)

What to do:

1. Review each workflow and its components.
2. Adjust failure modes, severity scores, and classifications as needed.
3. Add any failure modes the auto-import missed.
4. Verify criticality ratings (RPV ≥ 50 = High Risk).
5. Mark Phase 1 complete when satisfied.

4 - Phase 2: Mitigation & Validation

1. Navigate to Phase 2.
2. Click 🔄 Import from Phase 1 to pull in failure modes.
3. For each failure mode, review:
   • Mitigation strategy - detection, client-side, infrastructure, and recovery actions
   • Validation / Chaos test - the recommended test type (Chaos, Load, Failover, etc.)
4. Refine mitigations - the tool auto-suggests strategies based on Azure service type.
5. Download the XLSX export for offline review if needed.

5 - Phase 3: Health Model → Deploy to Azure

Phase 3 builds a hierarchical health model and deploys it.

1. Navigate to Phase 3.
2. Click 🔄 Import from Phase 1 to populate workflows, components, and FM-linked health signals.
3. Review the entity hierarchy:
   • Workflow → Business Function → Application Component → Platform Resource
4. For each entity, review health signals - metric name, thresholds, operators.
5. Add manual signals for any gaps; check the FMA Coverage score (aim for ≥ 80%).
6. When ready, click ☁️ Deploy to Azure on a workflow.
7. Fill in the deployment form:
   • Subscription ID
   • Resource Group
   • Location (Azure region)
   • Bearer Token (use Get-FoundryToken.ps1 or paste an existing token)
8. Click 🚀 Deploy Now - the tool creates entities and relationships in Azure Health Models.
9. On success, use the 🔗 Open in Azure Portal link to verify.

6 - Phase 4: Operations & Governance

After deployment, Phase 4 establishes ongoing operational practices.

1. Navigate to Phase 4.
2. Review the Operations Summary - open alerts, dependency alignment, auto-mitigation status.
3. Define the Governance Model:
   • Governance Area - what is being reviewed (e.g., "Health Signal Thresholds")
   • Owner - team or individual responsible
   • Frequency - review cadence (Weekly, Monthly, Quarterly)
   • Artifacts Reviewed - dashboards, runbooks, incident reports
   • Output - expected deliverable from each review cycle
4. Track the Governance Maturity Score (0–100) in the Insights panel - it scores FM coverage, mitigation completeness, signal coverage, and validation rate.

7 - Insights Dashboard

Click 📊 Insights at any time for a cross-phase summary:

• Workflow and component counts
• Failure mode coverage
• Signal-to-FM alignment
• Governance maturity score
• Export the full workbook (📥 Export All to XLSX) for stakeholder reporting

---

Data Persistence

All data is stored in the browser's localStorage. There is no server or database.

• Save - data auto-saves on every change
• Export - use per-phase or full XLSX exports to back up your work
• Import - XLSX upload restores previously exported data
• Clear - clearing browser data will erase all progress

Offline Usage

After the first load (which fetches CDN libraries), most features work offline. The exceptions are:

• AI Vision (Import Data Flows) - requires network access to the Azure OpenAI endpoint
• Deploy to Azure - requires network access to Azure Resource Manager
• Google Fonts - the app falls back to system fonts if unavailable