First Scan and Health Score

Run your first AutoDoctor scan, understand the current weighted health score model, and interpret findings, trends, and generated reports quickly.
Categories:
Tags:

Who This Is For

  • First-time users validating a successful installation.
  • Operators checking a host baseline.

Run the First Scan 

powershell -NoLogo -NoProfile -ExecutionPolicy Bypass -File "C:\ProgramData\AutoDoctor\agent\AutoDoctor.ps1"

If running from a cloned repo:

powershell -NoLogo -NoProfile -ExecutionPolicy Bypass -File ".\agent\AutoDoctor.ps1"

What Happens During a Scan

  1. AutoDoctor resolves paths, config, localization, and update metadata.
  2. The agent initializes SQLite and loads the module pipeline.
  3. Core modules collect system, memory, CPU, disk, network, event, startup, software, driver, and update data.
  4. Validation, history, anomaly, and correlation modules enrich the raw findings.
  5. Root Cause Analysis calculates the health score and summary.
  6. Automatic remediation runs in the full script.
  7. Diagnostics, telemetry, alerts, metadata, and reports are written.
Collection -> Validation/History/Anomaly/Correlation -> Root Cause Score -> Remediation -> DB/API/Reports

Health Score Logic (Current)

Score still starts at 100, but the current model is no longer a fixed seven-rule table.

AutoDoctor now combines:

  • direct threshold findings such as memory pressure, CPU saturation, disk pressure, high latency, and event errors
  • data quality findings from validation
  • anomaly and correlation findings
  • historical findings such as Sustained, Transient, and Baseline Deviation

The scoring flow is:

  1. Convert findings into ordered Critical, Warning, and Info items.
  2. Assign each finding a base weight.
  3. Apply a multiplier so repeated findings in the same category count less than the first one.
  4. Clamp the final score to the 0-100 range.

Practical effect:

  • One serious disk or memory problem can reduce the score quickly.
  • Repeated findings in the same domain still matter, but each additional one has less impact.
  • Trend-aware findings let AutoDoctor distinguish a one-off spike from a sustained problem.

Understanding the Trend Window

Historical scoring uses recent telemetry when it exists.

  • Preferred window: last 24 hours
  • Preferred sample depth: last 5 historical runs
  • Fallback behavior: if there are not enough samples in the time window, AutoDoctor falls back to the last few runs it can find

Common metric states:

  • Stable: current value is within normal range
  • Transient: current run crossed a threshold once
  • Sustained: threshold crossed across consecutive runs
  • Baseline Deviation: current value is materially worse than the rolling baseline
  • Increasing or Decreasing: directional trend without a threshold breach

Where to Read Results

  • HTML report: reports/AutoDoctor_Report.html
  • JSON report: reports/AutoDoctor_Report.json
  • Markdown report: reports/AutoDoctor_Report.md
  • PDF report: reports/AutoDoctor_Report.pdf when Chrome or Chromium is available
  • Meta file: server/latest_run.json
  • DB tables: diagnostics, alerts, telemetry_modules, system_info, telemetry_trends, telemetry_baselines, remediation

Quick Interpretation Tips

  • 90-100: no major issues detected or only low-impact findings
  • 70-89: moderate degradation, often one clear bottleneck or a few warnings
  • < 70: sustained or critical conditions are likely present and should be reviewed immediately

Start with these parts of the HTML report:

  • Why The Score Changed
  • Current State
  • Sustained Issues
  • Automatic Remediation

Next Steps