Performance Prophet

┌─────────────┐     ┌─────────────┐     ┌─────────────┐
│  Collectors │ ──▶ │  Predictor  │ ──▶ │   Alerter   │
└─────────────┘     └─────────────┘     └─────────────┘
     │                    │                    │
     ▼                    ▼                    ▼
  Metrics DB         Model State          Notifications

# Start collector daemon
prophet collect --interval 60

# Show current predictions
prophet predict

# Analyze specific metric
prophet analyze response_time --hours 24

# Show predicted incidents
prophet forecast --horizon 1h

# Interactive dashboard
prophet watch

# Pattern Learning (Gen 6+)
prophet patterns               # Show learned patterns & stats
prophet similar cpu_percent    # Find similar past incidents
prophet cascade memory_percent # Predict cascade effects
prophet link db_connections response_time 2.5 0.85  # Record causal link
prophet incident response_time --severity warning --cause "DB pool exhausted"

# Automatic Causality Discovery (Gen 6 Evolution)
prophet causality discover               # Auto-discover all causal relationships
prophet causality discover --metric cpu  # Discover relationships for specific metric
prophet causality explain --metric response_time  # Explain what causes/affects a metric
prophet causality graph                  # Generate Mermaid diagram of causal graph
prophet causality summary                # Summary of discovered relationships

# Causal Analysis
prophet causal-explain response_time     # What causes response_time? What does it affect?
prophet causal-roots error_rate          # Trace back to root causes
prophet causal-test memory_percent response_time  # Test if one metric causes another

# [home-path]
metrics:
  response_time:
    threshold: 500ms
    warning_horizon: 30m
    seasonality: daily

  memory_percent:
    threshold: 85
    warning_horizon: 15m

  error_rate:
    threshold: 0.05
    anomaly_sensitivity: high

notifications:
  slack: "#ops-alerts"
  email: [email]
  cooldown: 5m

from prophet import Prophet

p = Prophet()
p.track_session_metric("response_time", response_ms)

if p.predict_breach("response_time", horizon="30m"):
    alert("Slowdown predicted in next 30 minutes")

- [ ] Performance Prophet status
- [ ] Any predicted incidents?

🌊 Cascade Prediction: database_connections
  → response_time (lag: 2.5 min, correlation: 85%)
  → error_rate (lag: 5.0 min, correlation: 72%)

prophet similar response_time

🔍 Similar past incidents:
  📋 inc_20260131_response
     Root cause: Database connection pool exhausted
     Resolution: Increased pool size to 100
     Duration: 12.3 min

# Predict TTR for an active incident
prophet ttr response_time --value 650 --threshold 500 --elapsed 5 --trend stable

# Output:
# ⏱️  TTR Prediction: response_time
# ───────────────────────────────────
#    Estimated: 12 min remaining
#    Confidence: ████████░░ 80%
#
#    📊 Range:
#       Best case:  5 min
#       Most likely: 12 min
#       Worst case: 25 min
#
#    📋 Factors:
#       • Based on 8 similar past incidents
#       • Pattern type: gradual rise
#       • ⏱️ 5 min elapsed
#
#    💡 Recommendation:
#       Check: Database connection pool exhausted (common cause in similar incidents)

# View prediction accuracy
prophet ttr-stats

🔴 INCIDENT: response_time breached threshold (650ms > 500ms)
   ⏱️ Estimated recovery: 15 min (confidence: 75%)
   💡 Similar to 3 past incidents - common cause: DB connection pool

# Define an availability SLO
prophet slo-define api_availability \
  --target 0.999 \
  --window 30 \
  --metric response_success_rate \
  --policy "Freeze deployments, prioritize reliability"

# Define a latency SLO
prophet slo-define api_latency \
  --target 0.95 \
  --window 30 \
  --policy "No risky changes until budget recovers"

# List all SLOs
prophet slo-list

# Record good/total events (for event-based SLOs)
prophet slo-events api_availability 9990 10000

# Record a burn event with cause
prophet slo-burn api_availability 0.001 --cause "Deployment hiccup" --incident INC-123

# Show budget status for all SLOs
prophet slo-budget

# Output:
# 📊 SLO Budget Summary
#
#    🟡 api_availability
#       [████████████░░░] 85.2% remaining
#       Exhaustion in: 12.3d
#
#    🔴 api_latency
#       [████░░░░░░░░░░░] 28.1% remaining
#       Exhaustion in: 2.1d

# Detailed prediction for specific SLO
prophet slo-budget api_availability

# Output:
# 🟡 SLO Budget: api_availability
# ────────────────────────────────────────
#    Budget:     [████████████████░░░░] 85.2%
#    Status:     nominal
#    Confidence: 80%
#    Exhaustion: 12.3 days
#
#    📋 85.2% budget remaining. Burn rate is at expected.
#
#    💡 Recommendations:
#       • 📈 Monitor burn rate trend closely
#       • 🔧 Address known reliability issues

# Get JSON for dashboards
prophet slo-budget api_availability --json

prophet slo-alerts

# Output:
# 📊 api_availability
#    🚨 PAGE: 2.00% burned in 5m, 3.50% in 60m (threshold: 2%)
#
#    Burn rates:
#       5m:  0.0040%/min
#       1h:  0.0006%/min
#       1d:  0.0001%/min
#       Trend: rising

🔴 INCIDENT: api_latency breached p99 threshold
   ⏱️ Estimated recovery: 15 min
   💸 Projected budget burn: 0.3%
   📊 Budget remaining after: 27.8%

# [home-path]
slos:
  api_availability:
    target: 0.999
    window_days: 30
    policies:
      danger:
        - freeze_deployments
        - page_oncall
        - notify_stakeholders
      exhausted:
        - block_ci_merges
        - escalate_leadership

# Discover all causal relationships
prophet causality discover

# Output:
# 🧬 Discovering causal relationships...
#
# ✅ Discovered 5 causal links:
#
# 📊 memory_percent
#    → response_time
#       [████████░░] 82% strength
#       Lag: 2.5m | Method: granger
#    → disk_io
#       [██████░░░░] 61% strength
#       Lag: 5.0m | Method: cross_correlation

# Explain a specific metric
prophet causal-explain response_time

# Output:
# 🔍 Causal Analysis: response_time
#
# 📥 CAUSED BY:
#    ← memory_percent (strength: 82%, lag: 2.5m)
#    ← db_connections (strength: 71%, lag: 1.0m)
#
# 📤 CASCADES TO:
#    → error_rate (+3.0m, strength: 65%)
#       → user_complaints (+10.0m, strength: 45%)

# Test a specific relationship
prophet causal-test memory_percent response_time

# Output:
# 🔬 Testing: memory_percent → response_time
#
# 1️⃣ Cross-Correlation:
#    Peak lag: 2 minutes
#    Correlation: 0.823
#    Direction: memory_percent causes response_time
#
# 2️⃣ Granger Causality:
#    F-statistic: 15.32
#    Optimal lag: 2 minutes
#    Significance: strong
#    ✅ Granger causes: Yes
#
# 📋 Summary:
#    ✅ Evidence of causality: Correlated (0.82), Granger causes (F=15.3)

# Find root causes
prophet causal-roots error_rate --depth 3

# Output:
# 🌳 Root Causes: error_rate
#
# 📍 memory_percent (depth: 2)
# 📍 db_pool_size (depth: 2)
# 📍 traffic_load (depth: 1)

# Generate visual graph
prophet causality graph

# Output (Mermaid):
# ```mermaid
# graph LR
#     memory_percent--2m-->response_time
#     memory_percent--5m-->disk_io
#     response_time--3m-->error_rate
# ```

graph LR
    memory_percent--2m-->response_time
    memory_percent--5m-->disk_io
    db_connections--1m-->response_time
    response_time--3m-->error_rate
    error_rate--7m-->user_complaints

🔴 INCIDENT: error_rate breached threshold (5% > 1%)

🔍 Root Cause Analysis:
   └── memory_percent spike detected 8 minutes ago
       └── response_time degraded 5 minutes ago
           └── error_rate breached now

💡 Likely root cause: memory_percent (82% confidence)
   Historical pattern: Memory pressure → slow responses → errors

# Record a deployment
prophet event deployment "Deployed v2.3.1" --severity medium

# Record a config change
prophet event config_change "Increased connection pool to 100" --severity low

# Record a scheduled job
prophet event scheduled_job "Nightly batch processing" --impact-window 30

# Record traffic spike
prophet event traffic_spike "Marketing campaign launch" --severity high

# Collect deployments from git automatically
prophet event-collect git --path /path/to/repo --hours 24

# Attribute a current anomaly
prophet attribute response_time

# Output:
# 🔍 Attributing anomaly: response_time at 2024-01-15 14:30
#    Looking back 60 minutes for events...
#
# ## Anomaly Attribution Report
#
# ### 1. 🟢 Deployed v2.3.1
# - **Type:** deployment
# - **Confidence:** high (78%)
# - **Lag:** 12.3 minutes before anomaly
# - **Temporal fit:** 95%
# - **Metric relevance:** 100%
# - **Status:** Pending confirmation (ID: `a7b3c9d2`)
#
# ### 2. 🟡 Increased connection pool
# - **Type:** config_change
# - **Confidence:** medium (52%)
# - **Lag:** 45.1 minutes before anomaly

# Attribute a past anomaly
prophet attribute cpu_percent --timestamp 2024-01-15T10:00:00 --lookback 120

# Confirm an attribution was correct
prophet event-confirm a7b3c9d2 --yes

# Reject an incorrect attribution
prophet event-confirm b8c4d0e3 --no

# Show what Prophet has learned
prophet event-patterns

# Output:
# 🧠 Learned Attribution Patterns
#
# 📍 deployment → response_time
#    Typical lag: 8.5 min
#    Duration: 25.3 min
#    Occurrences: 12
#    Confidence boost: +35%
#
# 📍 scheduled_job → cpu_percent
#    Typical lag: 1.2 min
#    Duration: 15.0 min
#    Occurrences: 28
#    Confidence boost: +50%

prophet event-summary --hours 48

# Output:
# 📊 Attribution Summary (last 48h)
#
# By Confidence Level:
#    🟢 high: 15
#    🟡 medium: 8
#    🟠 low: 3
#
# By Event Type:
#    • deployment: 12
#    • scheduled_job: 8
#    • traffic_spike: 4
#
# Confirmation Status:
#    ✅ Confirmed: 18
#    ❌ Rejected: 3
#    ⏳ Pending: 5
#    📈 Accuracy: 86%
#
# 🧠 Learned Patterns: 6

🔴 INCIDENT: response_time breached threshold (850ms > 500ms)

🔍 Causal Analysis:
   Metric chain: memory_percent → response_time (lag: 2.5m)

🎯 Event Attribution:
   1. 🟢 [HIGH] Deployed v2.3.1 (12 min ago)
      Historical pattern: deployments → response_time (+15%, 8m lag)
   2. 🟡 [MEDIUM] DB pool config change (45 min ago)

💡 Likely root cause: Deployment v2.3.1
   Suggested action: Review recent deployment changes

# Generate capacity plan
prophet capacity

# Output:
# 📊 Capacity Planning Report
#    Generated: 2024-01-15 14:30
#    Horizon: 90 days
#
# ✅ All systems have comfortable capacity headroom
#    Risk Score: 10%
#
# 📈 Capacity Forecasts:
#
# 🟢 cpu_percent
#    Utilization: [████░░░░░░] 42%
#    Current: 42.3 / 85 (threshold)
#    Trend: stable/declining ✓
#    Growth: -0.12/day (stable)
#    Confidence: 78%
#
# 🟡 disk_percent
#    Utilization: [███████░░░] 68%
#    Current: 68.2 / 90 (threshold)
#    Time to breach: 2.3 weeks
#    Forecast date: 2024-01-29
#    Growth: +1.35/day (linear)
#    Confidence: 85%
#
# 🔧 Scaling Recommendations:
#
# 🟡 disk_percent
#    Action: plan_scaling
#    Plan disk_percent capacity increase for ~2.3 weeks
#    Current: 68.2 → Recommended: 115.0
#    Act by: 2024-01-22
#    Cost impact: ~1.2x
#    • Forecast breach: 2024-01-29
#    • Time to plan and budget
#    Alternatives:
#      - Monitor and reassess in 1 week

# Full capacity plan with all metrics
prophet capacity

# Detailed forecast for specific metric
prophet capacity-forecast disk_percent

# Analyze growth patterns
prophet capacity-growth memory_percent --window 14

# View/set thresholds
prophet capacity-threshold
prophet capacity-threshold cpu_percent --warning 70 --critical 85 --max 100

# Record scaling actions (for learning)
prophet capacity-scale disk_percent scale_up --old 100 --new 200 --description "Added second SSD"

# View scaling history
prophet capacity-history --days 30
prophet capacity-history --metric disk_percent

# Compare forecasts over time
prophet capacity-compare disk_percent --days 7

prophet capacity-growth memory_percent

# Output:
# 📊 Growth Analysis: memory_percent
#
#    Current value: 72.50
#    Daily growth: +0.85
#    Growth rate: +1.17%/day
#
#    Trend: linear
#    Seasonality: daily
#    Volatility: 0.15
#
#    Data points: 2048
#    Window: 14 days
#    Confidence: 82%
#
# Projections (at current rate):
#    7d: 78.45
#    30d: 98.00
#    90d: 149.00

# After upgrading disk
prophet capacity-scale disk_percent scale_up \
  --old 100 --new 200 \
  --description "Upgraded to 200GB SSD"

# After adding instances
prophet capacity-scale cpu_percent scale_out \
  --old 100 --new 200 \
  --description "Added second server"

# After optimization
prophet capacity-scale memory_percent optimize \
  --old 85 --new 60 \
  --description "Fixed memory leak in API service"

prophet capacity-compare disk_percent --days 7

# Output:
# 📈 Forecast Trend: disk_percent (last 7 days)
#
# Date               Days to Threshold    Growth Rate     Confidence
# ----------------------------------------------------------------------
# 2024-01-08 10:30   25.3 days           +0.98/day       75%
# 2024-01-09 10:30   22.1 days           +1.12/day       78%
# 2024-01-10 10:30   19.8 days           +1.25/day       80%
# 2024-01-11 10:30   16.2 days           +1.35/day       82%
# 2024-01-12 10:30   14.5 days           +1.35/day       85%

# Production: stricter thresholds
prophet capacity-threshold cpu_percent --warning 60 --critical 75 --max 100
prophet capacity-threshold memory_percent --warning 70 --critical 85 --max 100

# Development: more relaxed
prophet capacity-threshold disk_percent --warning 85 --critical 95 --max 100

## Daily Capacity Check
- [ ] Run `prophet capacity` — any urgent scaling needed?
- [ ] Critical resources: < 7 days to threshold → escalate
- [ ] Planning horizon: 14-30 days → add to sprint backlog

# Cost model in config
[home-path]
{
  "cost_model": {
    "cpu": {"scale_factor": 2, "cost_ratio": "2x"},
    "memory": {"scale_factor": 2, "cost_ratio": "1.5x"},
    "disk": {"scale_factor": 2, "cost_ratio": "1.2x"}
  }
}

# Get remediation recommendations for a metric issue
prophet remediate response_time --value 850 --threshold 500

# Output:
# 🔧 Remediation Recommendations
#
# ### 1. Rollback recent deployment
#    [████████░░] 78% confidence
#    Revert to previous known-good version
#
#    📊 Expected improvement: 90%
#    ⏱️  Time to effect: ~5 min
#    🟢 Effort: low
#    🟡 Risk: medium
#    📋 Based on 12 past uses (83% success)
#
#    💡 Evidence:
#       • Recent deployment detected: v2.3.1
#       • Historical success rate: 83%
#       • Recommended for 'spike' pattern
#
# ### 2. Increase connection pool
#    [██████░░░░] 65% confidence
#    Increase database connection pool size
#    ...

# Get recommendations for any metric
prophet remediate memory_percent --value 92 --threshold 85

# Include pattern context for better recommendations
prophet remediate cpu_percent --value 95 --pattern gradual_rise

# Skip causal analysis (faster, but less accurate)
prophet remediate response_time --value 800 --no-causality

# Record a remediation outcome (for learning)
prophet remediate-record mem_restart_service memory_percent \
  --before 92 --after 45 --threshold 85 --success \
  --time-to-effect 2 --duration 480

# View remediation statistics
prophet remediate-stats
prophet remediate-stats --action-id mem_restart_service

# Add a custom remediation action
prophet remediate-action --add \
  --name "Restart API Gateway" \
  --description "Restart the API gateway to clear stale connections" \
  --category restart \
  --metrics response_time,error_rate \
  --commands "kubectl rollout restart deployment/api-gateway" \
  --impact 75 \
  --effort low \
  --risk low \
  --tags quick_fix,gateway

# List custom actions
prophet remediate-action
prophet remediate-action --metric response_time

# Create a playbook for memory issues
prophet remediate-playbook --create \
  --name "Memory Pressure Response" \
  --description "Standard response for memory exhaustion" \
  --trigger-metric memory_percent \
  --trigger-pattern gradual_rise \
  --actions "mem_restart_service,mem_increase_limit,mem_fix_leak"

# Find matching playbook for a situation
prophet remediate-playbook --match memory_percent --pattern gradual_rise

# List all playbooks
prophet remediate-playbook

🔴 INCIDENT: response_time breached threshold (850ms > 500ms)

🔍 Root Cause Analysis:
   └── db_connections spike detected 5 minutes ago

🔧 Recommended Actions:
   1. [78%] Increase connection pool (effort: trivial)
   2. [65%] Rollback recent deployment (effort: low)
   3. [52%] Add read replica (effort: high)

💡 Start with: Increase connection pool
   Highest confidence, lowest effort, addresses root cause

# After successful remediation
prophet remediate-record db_increase_pool db_connections \
  --before 95 --after 40 --threshold 80 --success \
  --time-to-effect 1 --duration 1440 \
  --notes "Increased pool from 50 to 100"

# After failed remediation
prophet remediate-record rt_add_cache response_time \
  --before 850 --after 800 --threshold 500 \
  --notes "Cache helped but not enough - need DB optimization"

# Explain a current prediction
prophet explain response_time

# Demo with sample data
prophet explain-demo

# Show what changes would help
prophet counterfactual response_time

# Show feature contributions
prophet contributors response_time

# Update baselines from history
prophet update-baselines --days 7

## 🟠 Prediction Explanation

**Metric:** response_time
**Predicted Value:** 650.00
**Confidence:** 85%
**Severity:** WARNING

### Summary
Response times are predicted to reach 650ms within 25 minutes, which would
exceed your 500ms threshold. Users may start noticing delays.

### Why This Is Happening

- ↑ **cpu_percent**: above CPU usage (78.5% vs normal 45.2%) is contributing to the prediction
- ↑ **memory_percent**: above memory (82.3% vs normal 55.1%) contributing to slowdowns
- ↑ **queue_depth**: Queue depth of 156 (above from 45) contributing to delays

### Pattern Matches

- 🔍 Resource exhaustion pattern: Both CPU and memory elevated

### Recommended Actions

🟢 Profile and optimize CPU-intensive operations — Expected impact: 35% risk reduction
🟢 Tune garbage collection parameters — Expected impact: 28% risk reduction
🟡 Add more compute capacity or scale horizontally — Expected impact: 20% risk reduction

### What Would Help

- Reducing **cpu_percent** by 25% would reduce risk by ~35%
- Reducing **memory_percent** by 20% would reduce risk by ~28%
- Reducing **queue_depth** by 50% would reduce risk by ~20%

🔄 Counterfactuals for response_time
   Predicted: 650.00 → Threshold: 500

What would help:

1. 🟢 cpu_percent
   Current: 78.50 → Target: 58.87
   Change needed: 25% reduction
   Impact: ~35% risk reduction
   Action: Profile and optimize CPU-intensive operations

2. 🟢 memory_percent
   Current: 82.30 → Target: 65.84
   Change needed: 20% reduction
   Impact: ~28% risk reduction
   Action: Tune garbage collection parameters

# Update baselines from last week
prophet update-baselines --days 7

# Baselines are used to:
# - Calculate deviation (z-score) for each metric
# - Determine if values are unusually high/low
# - Weight contribution calculations

# Analyze a deployment (auto-detects git changes)
prophet deploy-risk

# Analyze specific files/changes
prophet deploy-risk --files "src/db/migrations/001.sql,src/api/routes.py" \
    --message "feat: add payment processing"

# With test coverage
prophet deploy-risk --coverage 75

# In CI/CD - fail if high risk
prophet deploy-risk --fail-on-high

# Record deployment for learning
prophet deploy-risk --record

# Record outcome after deployment
prophet deploy-record <deployment-id> --incident-24h --rollback

# View deployment history
prophet deploy-history --days 30

# Show learned patterns
prophet deploy-patterns

🔍 Analyzing deployment risk...

🟠 **Deployment Risk: 67/100 (HIGH)**

⚠️ **Recommendation: PROCEED WITH CAUTION** - Review risks below.

**Key Risk Factors:**
- Database migration detected
- 3 critical path files modified: src/db/migrations/001.sql, src/payments/handler.py
- Deploying on Friday - weekend/end of week
- Test coverage is 65% (below 70% threshold)

### Signal Breakdown

| Signal | Category | Value | Contribution |
|--------|----------|-------|--------------|
| migration_detected | code | 90% | ████ 0.23 |
| hotpath_touched | code | 60% | ██ 0.13 |
| deploy_day_risk | timing | 50% | █ 0.06 |
| test_coverage_delta | context | 14% | 0.03 |

### Recommendations

1. 🚨 Consider delaying this deployment
2. Test migration on production copy first
3. Ensure backward compatibility
4. Have rollback migration ready
5. Deploy during low-traffic window

**Optimal Deploy Window:** Tuesday 10:00 AM (lower risk window)

### Similar Past Deployments

- 2026-01-15: risk 72 → 🔴 incident (rolled back)
- 2026-01-08: risk 45 → ✅ ok

# GitHub Actions example
deploy:
  steps:
    - name: Analyze deployment risk
      run: |
        prophet deploy-risk --fail-on-high --record

    - name: Deploy (if safe)
      if: success()
      run: ./deploy.sh

    - name: Record outcome
      if: always()
      run: |
        if [ "${{ job.status }}" == "failure" ]; then
          prophet deploy-record $DEPLOYMENT_ID --incident-24h
        fi

# After successful deployment
prophet deploy-record abc123def456

# After incident within 24h
prophet deploy-record abc123def456 --incident-24h

# After rollback
prophet deploy-record abc123def456 --rollback --incident-1h

prophet deploy-patterns

# Output:
# 📊 Learned Deployment Patterns
#
# **Changes to: src/db, src/api**
#    🔴 HIGH RISK
#    Deployments: 15 | Incidents: 5 (33%)
#    Avg Risk Score: 68
#
# **Changes to: src/utils**
#    🟢 SAFE
#    Deployments: 42 | Incidents: 0 (0%)
#    Avg Risk Score: 22

# View adaptive threshold status
prophet adaptive status

# Register a metric with base thresholds
prophet adaptive register cpu_percent 75 90 above
prophet adaptive register response_time 500 1000 above

# Learn optimal thresholds from history
prophet adaptive learn cpu_percent

# Check a value against adaptive thresholds
prophet adaptive check cpu_percent 82

# View growth trends
prophet adaptive growth

# Starting a deployment (relax thresholds temporarily)
prophet adaptive context start deployment

# Deployment complete
prophet adaptive context end deployment

# Maintenance window
prophet adaptive context start maintenance
prophet adaptive context end maintenance

# View recent violations
prophet adaptive violations

# Output:
# [42] cpu_percent = 78.50 (warning) @ 2025-01-15 10:30 [?]
# [41] memory_percent = 85.20 (warning) @ 2025-01-15 09:15 [?]
# [40] response_time = 520.00 (warning) @ 2025-01-14 14:22 [✓]

# Provide feedback: was it a valid alert?
prophet adaptive feedback 42 invalid   # False positive → relax thresholds
prophet adaptive feedback 41 valid     # Good catch → maintain thresholds
prophet adaptive feedback 40 valid 15  # Valid, resolved in 15 minutes

prophet adaptive suggest

# Output:
# 🔴 [relax_thresholds] cpu_percent
#    High false positive rate (45%)
#    Suggest: warning=86.25, critical=103.50
#
# 📈 [growth_warning] memory_percent
#    Gradual growth (18.5%)
#    🚨 memory_percent has grown 18.5% over analyzed period.
#    Consider adjusting base thresholds upward.

prophet adaptive growth

# Output:
# 📈 cpu_percent: growing (12.3%)
# ➡️ memory_percent: stable (-2.1%)
# 🚀 disk_usage: rapid_growth (35.7%)

cpu_percent thresholds:
  peak:     warning=82, critical=95 (learned, 87% confidence)
  off_peak: warning=70, critical=88 (learned, 72% confidence)
  quiet:    warning=45, critical=70 (learned, 65% confidence)
  weekend:  warning=55, critical=80 (learned, 58% confidence)

prophet adaptive status

# Output:
# 🎯 Adaptive Thresholds Status
#
# **Metrics Tracked:** 5
# **Metrics Adapted:** 3
# **Violations (30d):** 47
# **False Positive Rate:** 18.5%
# **Pending Feedback:** 12
#
# ## Per-Metric Status
#
# ### 🟢 cpu_percent (✓ adapted)
# - Warning: 75.00 → **82.50**
# - Critical: 90.00 → **97.35**
# - Context: peak | Confidence: 87%
# - Violations: 12 | FP: 2 | Pending: 1

# List all fingerprints
prophet fingerprint list

# Capture current incident as a fingerprint
prophet fingerprint capture "Memory leak in API" --resolution "Restart pods" --tags memory,api

# Match current anomalies against library
prophet fingerprint match

# Show fingerprint details
prophet fingerprint show fp_abc123

# Update a fingerprint
prophet fingerprint update fp_abc123 --resolution "Restart pods AND fix query in /api/heavy"

# View library statistics
prophet fingerprint stats

# Basic capture
prophet fingerprint capture "DB connection exhaustion"

# With resolution and metadata
prophet fingerprint capture "Cache miss storm" \
  --lookback 45 \
  --resolution "Increase Redis maxmemory, add cache warming" \
  --tags cache,redis,performance \
  --severity critical

prophet fingerprint match

# Output:
# 🔍 Found 2 matching fingerprint(s)
#
# 🟢 **Memory leak in API service** (87% match)
#    Matched: memory_percent, response_time, gc_pause_ms
#    Missing: none
#    💡 Suggested: Restart API pods, investigate /api/heavy-query
#    Last seen: 12d ago
#
# 🟡 **Cache invalidation cascade** (64% match)
#    Matched: memory_percent, response_time
#    Missing: cache_hit_rate
#    Extra: gc_pause_ms
#    💡 Suggested: Check cache TTL settings
#    Last seen: 23d ago

# Show detailed fingerprint
prophet fingerprint show fp_abc123

# Update metadata
prophet fingerprint update fp_abc123 \
  --resolution "Better fix: Add connection pooling" \
  --ttr 25 \
  --notes "Happens after deployments with DB migrations"

# Delete a fingerprint
prophet fingerprint delete fp_abc123

# Find similar fingerprints (merge candidates)
prophet fingerprint similar fp_abc123

# Merge similar fingerprints
prophet fingerprint merge fp_abc123 fp_def456 --name "API memory issues (merged)"

# When a match was helpful
prophet fingerprint feedback fp_abc123 correct

# When a match was wrong
prophet fingerprint feedback fp_abc123 incorrect "Different root cause - was network latency"

prophet fingerprint stats

# Output:
# 📊 Fingerprint Library Statistics
#
#    Total fingerprints: 24
#    Unique metrics tracked: 8
#    Total occurrences: 156
#    Average TTR: 18.5 minutes
#
#    By severity:
#      🔵 info: 5
#      🟡 warning: 14
#      🔴 critical: 5
#
#    Most recurring patterns:
#      • Memory leak in API: 23 matches
#      • DB connection pool exhaustion: 18 matches
#      • Cache miss storm: 12 matches

prophet fingerprint show fp_abc123

# Output includes:
# 📋 Resolution Playbook:
#    Resolution: Restart API pods, investigate /api/heavy-query
#    Expected TTR: 15 minutes
#    Occurrences: 23
#    Match accuracy: 87% (19 matches)

# In HEARTBEAT.md or monitoring:
prophet fingerprint match --json | jq '.matches[0].suggested_resolution'

# List available scenarios
prophet simulate list

# Run a preset scenario
prophet simulate run traffic_2x
prophet simulate run server_loss
prophet simulate run memory_leak

# Quick what-if analysis
prophet simulate what-if memory_percent 95
prophet simulate what-if cpu_percent 90

# Find breaking points
prophet simulate breaking-point cpu_percent
prophet simulate breaking-point memory_percent

# Compare scenarios (which is safer?)
prophet simulate compare traffic_2x server_loss

============================================================
SIMULATION: Double Traffic
============================================================

🟠 Impact: SIGNIFICANT
📊 Risk Score: 84/100
⏱️  Duration: 60 minutes

🚨 Threshold Breaches:
   • response_time: 520.0 (threshold: 500.0)
   • request_rate: 10000.0 (threshold: 10000.0)

🔗 Cascade Effects:
   cpu_percent → response_time
   response_time → error_rate
   queue_depth → response_time

📈 SLO Impact:
   latency_p99: █████░░░░░ 52%
   error_budget: ░░░░░░░░░░ 3%

⚠️  Warnings:
   ⚠️ Deep cascade detected (3 levels) - failure can propagate quickly

💡 Recommended Mitigations:
   1. Pre-scale horizontally before expected traffic increase
   2. Enable auto-scaling with appropriate thresholds
   3. Add response caching for common requests

prophet simulate breaking-point cpu_percent

# Output:
# 📊 Breaking Point Analysis: cpu_percent
#    Baseline value: 42.5
#    ⚠️  Breaks at: 2.5x (106.3)
#    📏 Headroom: 150%

prophet simulate compare traffic_2x scale_up_2x chaos_mild

# Output:
# 🟢 SAFEST: Double Capacity
#    Risk Score: 15/100
#    Impact: minimal
#
# 🔴 RISKIEST: Double Traffic
#    Risk Score: 84/100
#    Impact: significant
#
# All Results (sorted by risk):
#    1. 🟢 Double Capacity: 15/100 (minimal)
#    2. 🟡 Mild Chaos: 45/100 (moderate)
#    3. 🔴 Double Traffic: 84/100 (significant)

# Before expected traffic event
prophet simulate run traffic_5x

# If risk is high, test scaling
prophet simulate run scale_up_2x
prophet simulate compare traffic_5x scale_up_2x

prophet simulate custom \
  --name "Peak Hour Load" \
  --metric cpu_percent 85 \
  --metric memory_percent 80 \
  --duration 120

Technical: response_time = 2500ms
    ↓
User: "Frustrating - attention wanders"
    ↓
Business: ~70 abandonments/hr, $4,350/hr revenue loss

# System-wide impact summary
prophet impact

# List all configured user journeys
prophet impact-journey

# Impact for specific journey
prophet impact-journey checkout

# Impact of a specific metric
prophet impact-metric response_time --value 2500 --journey checkout

# Add or update a user journey
prophet impact-add --name payment \
  --stage CONVERSION \
  --traffic 500 \
  --conversion-rate 0.8 \
  --revenue 120 \
  --latency-target 300 \
  --latency-sensitivity 2.0

# Show impact trend over time
prophet impact-trend
prophet impact-trend checkout --hours 48

Response Time → Abandonment
     0.5s    →    1%
     1.0s    →    3%
     2.0s    →    7%
     3.0s    →   13%
     4.0s    →   20%
     5.0s    →   28%
     6.0s    →   35%
     8.0s    →   45%
    10.0s    →   53%

╔════════════════════════════════════════════╗
║     👤 USER EXPERIENCE IMPACT SUMMARY      ║
╚════════════════════════════════════════════╝

UX Score: 🟡🟡🟡⚪⚪ 60/100 (Degraded)
Frustration Index: 40/100

👥 Users Affected: ~50,000/hr
🚪 Max Abandonment: 3.0%

💸 Hourly Revenue Impact: $1,894
📊 Projected Daily Loss: $45,465

📈 Predicted UX Score: 55/100
⚠️  Critical in: ~2.0 hours

🔻 Most Affected Journeys:
   • checkout: 60/100 health
   • search: 66/100 health
   • api: 71/100 health

🚨 Critical Issues:
   • response_time on checkout: 40 frustration

🔴 **checkout** — SEVERE
   Metric: response_time = 2500.0
   Frustration Score: 70/100

   👥 Affected Users: ~1,000/hr
   🚪 Abandonment Rate: 7.0%
   ❌ Est. Abandonments: ~70/hr

   💸 Lost Conversions: ~58/hr
   💰 Revenue Impact: $4,350/hr

   🎯 User Experience: Response: Frustrating - attention wanders

# High-value subscription flow
prophet impact-add --name subscription_checkout \
  --stage CONVERSION \
  --traffic 200 \
  --conversion-rate 0.4 \
  --revenue 299 \
  --latency-target 250 \
  --latency-sensitivity 2.5 \
  --error-sensitivity 3.0 \
  --metrics "payment_latency,subscription_api"

# View memory statistics
prophet memory stats

# Recent incidents
prophet memory recent [limit]

# Search incidents by keyword
prophet memory search "database"

# Record a new incident (interactive)
prophet memory record

# Resolve an incident
prophet memory resolve <incident_id> "Increased connection pool"

# Check for predicted incidents based on current state
prophet memory predict

# View learnings (root causes, remediations, preventions)
prophet memory learnings [incident_id]

# Add a learning to an incident
prophet memory add-learning <incident_id> remediation "Restart the worker pods"

# View recurring patterns
prophet memory recurring

# Auto-record from current breached metrics
prophet memory-auto --title "API slowdown"

⚠️  Memory exhaustion during batch processing
   Type: daily (every 24.0h)
   Next predicted: 2024-02-03 03:00:00
   Confidence: 78%

⚠️  INCIDENT PREDICTION

Category: saturation
Severity: sev2
Probability: 72%

Warning Signs:
  • Similar to Database connection exhaustion: metrics: memory_percent, db_connections
  • memory_percent pattern: gradual_rise
  • Recurring pattern: occurs daily, next expected 2024-02-03 03:00

Similar Past Incidents:
  • Database connection exhaustion (78% match)
  • Memory pressure incident (62% match)

Suggested Actions:
  ✓ Increase connection pool size before peak
  ✓ Pre-emptive: Restart worker pods
  ✓ Enable connection pool monitoring

📚 LEARNINGS FROM SIMILAR INCIDENTS

Root Causes (most common):
  → Connection pool exhausted under load (seen 5x)
  → Batch job holding connections (seen 3x)

Effective Remediations:
  ✓ Increase pool size to 100 (85% effective)
  ✓ Add connection timeout (78% effective)

Prevention Measures:
  🛡 Monitor connection pool utilization
  🛡 Set alerts at 70% pool usage

# Show overall calibration statistics
prophet calibration stats

# Output:
# ╔════════════════════════════════════════════════════════════╗
# ║              📊 CALIBRATION STATISTICS                     ║
# ╚════════════════════════════════════════════════════════════╝
#
# 📝 Total Predictions Tracked: 247
# ⏳ Pending Evaluation: 12
# ✅ Evaluated: 235
#
# 🎯 Overall Accuracy: 78.3%
#    Correct: 184 | Incorrect: 51
# 📊 Average Confidence: 76.1%
# 📈 Recent 7-Day Accuracy: 81.2%

# Detailed calibration report
prophet calibration report

# Output:
# ╔════════════════════════════════════════════════════════════╗
# ║            🎯 CONFIDENCE CALIBRATION REPORT               ║
# ╚════════════════════════════════════════════════════════════╝
#
# ┌─────────────────────────────────────────────────────────┐
# │ CALIBRATION BY CONFIDENCE BUCKET                        │
# ├────────────┬───────┬─────────┬──────────┬───────────────┤
# │ Confidence │ Count │ Correct │ Accuracy │ Calibration   │
# ├────────────┼───────┼─────────┼──────────┼───────────────┤
# │   0%-10%   │     5 │       0 │       0% │ 🎯 Excellent  │
# │  10%-20%   │     8 │       2 │      25% │ ✅ Good       │
# │  20%-30%   │    12 │       3 │      25% │ ⚠️ Fair       │
# │  30%-40%   │    18 │       8 │      44% │ ✅ Good       │
# │  40%-50%   │    22 │      11 │      50% │ 🎯 Excellent  │
# │  50%-60%   │    35 │      21 │      60% │ ✅ Good       │
# │  60%-70%   │    42 │      27 │      64% │ ⚠️ Fair       │
# │  70%-80%   │    48 │      39 │      81% │ ✅ Good       │
# │  80%-90%   │    31 │      26 │      84% │ 🎯 Excellent  │
# │  90%-100%  │    14 │      13 │      93% │ 🎯 Excellent  │
# └────────────┴───────┴─────────┴──────────┴───────────────┘

# Filter by prediction type
prophet calibration report --type threshold_breach --days 60

# Show predictions awaiting outcome evaluation
prophet calibration pending

# Manually record an outcome
prophet calibration resolve pred_20260202_abc123 correct 512.5 "Threshold was breached as predicted"

# Update adjustment factors (recalculate from outcomes)
prophet calibration update

# Show reliability by prediction type
prophet calibration reliability

# Track a prediction manually (for testing)
prophet calibration track threshold_breach response_time 0.85 30

from calibration import CalibrationEngine, CalibratedPredictor, PredictionType

# Wrap any predictor with calibration
engine = CalibrationEngine()
predictor = CalibratedPredictor(engine)

# Make a calibrated prediction
# Raw confidence 0.85 might become 0.78 based on historical accuracy
adjusted_conf = predictor.predict_threshold_breach(
    metric="response_time",
    raw_confidence=0.85,
    horizon_minutes=30,
    threshold=500
)
# Prediction is automatically tracked for later evaluation

from calibration import CalibrationEngine

engine = CalibrationEngine()

# Auto-evaluate threshold breach predictions
def get_metric_value(metric):
    return current_metrics[metric]

engine.auto_evaluate_threshold_predictions(get_metric_value)

# Auto-evaluate trend predictions
def get_metric_history(metric, hours):
    return historical_values[metric][-hours*60:]

engine.auto_evaluate_trend_predictions(get_metric_history)

# Make ensemble prediction
prophet ensemble predict cpu_percent
prophet ensemble predict response_time --horizon 60

# Show model performance statistics
prophet ensemble stats
prophet ensemble stats response_time  # For specific metric

# Show learned weights for metric
prophet ensemble weights cpu_percent

# Update weights from historical accuracy
prophet ensemble update-weights memory_percent

# Analyze when/why models disagree
prophet ensemble disagreement error_rate

# Predict threshold breach probability
prophet ensemble breach memory_percent 85
prophet ensemble breach cpu_percent 90 --horizon 30

# Show metric's pattern classification
prophet ensemble pattern response_time

╔════════════════════════════════════════════════════════════╗
║                🎯 ENSEMBLE PREDICTION                       ║
╚════════════════════════════════════════════════════════════╝

  Metric:      cpu_percent
  Current:     68.42
  Horizon:     30 steps
  Predicted:   72.15
  Confidence:  78.3%
  Uncertainty: 15.2%
  Agreement:   84.8%
  Dominant:    holt_winters

  ┌─────────────────────┬────────────┬────────────┬──────────┐
  │ Model               │ Prediction │ Confidence │ Weight   │
  ├─────────────────────┼────────────┼────────────┼──────────┤
  │ holt_winters        │      72.34 │      82.5% │   35.2%  │
  │ arima_lite          │      71.89 │      76.1% │   28.1%  │
  │ threshold_proj      │      73.12 │      71.3% │   18.5%  │
  │ percentile          │      71.45 │      68.9% │   12.1%  │
  │ seasonal_naive      │      72.00 │      55.0% │    6.1%  │
  └─────────────────────┴────────────┴────────────┴──────────┘

  ✅ High agreement — models converge on similar prediction

# After enough predictions with recorded outcomes...
prophet ensemble update-weights cpu_percent

✅ Updated weights for cpu_percent:

  holt_winters: 38.2%   # Best performer
  arima_lite: 26.1%
  threshold_proj: 17.4%
  percentile: 11.8%
  seasonal_naive: 6.5%  # Worst performer

Agreement: 42.3%  ❌ Low agreement — significant uncertainty

💡 Models diverging suggests:
   - Regime change in data
   - Unusual pattern not seen before
   - Higher risk of prediction error

prophet ensemble breach memory_percent 85

╔════════════════════════════════════════════════════════════╗
║              🚨 THRESHOLD BREACH PREDICTION                 ║
╚════════════════════════════════════════════════════════════╝

  Metric:       memory_percent
  Threshold:    85.0
  Current:      78.23
  Predicted:    82.45
  Horizon:      30 steps

  🟡 Breach Probability: 45.2% (LOW RISK)
  Confidence:   72.1%
  Agreement:    81.3%
  Time to Breach: ~42.3 steps

from ensemble import EnsemblePredictor, ThresholdBreachPredictor

# Basic ensemble prediction
ensemble = EnsemblePredictor()
prediction = ensemble.predict(
    metric="response_time",
    values=[100, 105, 110, 108, 115, 120],
    horizon=10
)

print(f"Predicted: {prediction.final_value}")
print(f"Confidence: {prediction.confidence:.1%}")
print(f"Agreement: {prediction.agreement_score:.1%}")

# Threshold breach prediction
breach_predictor = ThresholdBreachPredictor(ensemble)
result = breach_predictor.predict_breach(
    metric="memory_percent",
    values=memory_values,
    threshold=85.0,
    horizon_minutes=30
)

if result['breach_probability'] > 0.7:
    alert("High probability of memory breach!")

# Record outcomes for weight learning
ensemble.record_outcome(
    metric="response_time",
    prediction_type=PredictionType.VALUE_FORECAST,
    model_predictions=prediction.model_predictions,
    actual_value=118.5  # What actually happened
)

# Update weights periodically
ensemble.update_weights("response_time", PredictionType.VALUE_FORECAST)

# Current health score
prophet health

# Health score with JSON output
prophet health --json

# Health score trend (sparkline)
prophet health-history
prophet health-history --hours 48

# Auto-generated reports
prophet report              # Daily report (default)
prophet report hourly       # Last hour
prophet report daily        # Last 24 hours
prophet report weekly       # Last 7 days
prophet report daily --json # JSON output

✨ System Health: 95.5/100 (A)
   Trend: 📈 Improving

   🟡 calibration   75.0/100 — Building calibration data (6 predictions tracked)
   🟢 pattern       92.0/100 — No anomalous patterns detected
   🟢 incident      95.0/100 — No incidents in the last 24 hours
   🟢 prediction   100.0/100 — All 6 metrics within expected bounds
   🟢 threshold    100.0/100 — All 3 adaptive thresholds healthy
   🟢 capacity     100.0/100 — Healthy capacity: 100% avg headroom

   Recommendations:
   🟡 Monitor prediction accuracy — slight drift detected

   System is running like a well-oiled machine. ✨

══════════════════════════════════════════════════
  PERFORMANCE PROPHET — DAILY REPORT
  Generated: 2026-02-02 23:29
══════════════════════════════════════════════════

  ✨ Overall Health: 95.5/100 (A)
  📈 Period: avg 90 | low 78 | high 96

  📈 Trend: improving

  DIMENSIONS:
  ──────────────────────────────────────────────
  calibration  [▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓·····]  75.0
  pattern      [██████████████████··]  92.0
  incident     [███████████████████·]  95.0
  prediction   [████████████████████] 100.0
  threshold    [████████████████████] 100.0
  capacity     [████████████████████] 100.0

  ✅ IMPROVEMENTS:
    • threshold improved by 30 points
    • capacity improved by 30 points

  📋 RECOMMENDATIONS:
    🟡 Monitor prediction accuracy — slight drift detected

══════════════════════════════════════════════════

from health_score import HealthScoreEngine
engine = HealthScoreEngine()
snapshot = engine.calculate_snapshot()
if snapshot.overall_score < 70:
    # Alert: system health degraded
    notify(snapshot.narrative)

┌─────────────────────────────────────────────────────────┐
│                 All Prophet Modules                       │
│  Prophet │ Thresholds │ Calibration │ Fingerprints │ ... │
└────┬─────┴─────┬──────┴──────┬──────┴──────┬───────┴────┘
     │           │             │             │
     ▼           ▼             ▼             ▼
┌─────────────────────────────────────────────────────────┐
│              Health Score Engine                         │
│  Weighted combination → single 0-100 score              │
│  Grade assignment (A+ through F)                         │
│  Trend detection (improving/stable/degrading/volatile)  │
└────────────────────┬────────────────────────────────────┘
                     │
                     ▼
┌─────────────────────────────────────────────────────────┐
│              Narrative Generator                         │
│  Human-readable stories from numbers                    │
│  Top concerns + strengths + recommendations             │
└────────────────────┬────────────────────────────────────┘
                     │
                     ▼
┌─────────────────────────────────────────────────────────┐
│              Auto-Report Engine                          │
│  Hourly / Daily / Weekly reports                        │
│  Score history + sparklines + trend analysis            │
│  Period comparisons + notable events                    │
└─────────────────────────────────────────────────────────┘

# Comprehensive drift check
prophet drift                    # Full drift report with trust scores
prophet drift --metric cpu       # Check specific metric
prophet drift --json             # JSON output

# Trust scores
prophet drift-trust              # All metric trust scores
prophet drift-trust cpu          # Trust for specific metric

# Operational regime
prophet drift-regime             # Current regime + transitions
prophet drift-regime --json      # JSON output

# Signal history
prophet drift-history            # Last 24h signals
prophet drift-history --hours 72 # Custom lookback
prophet drift-history --severity severe  # Filter by severity

# Prediction accuracy
prophet drift-accuracy           # All metrics
prophet drift-accuracy cpu       # Specific metric

# Resolve signals
prophet drift-resolve 42 --note "Deployed model v2"

# Interactive demos
prophet drift-demo --scenario gradual   # Slow mean shift
prophet drift-demo --scenario sudden    # Abrupt distribution change
prophet drift-demo --scenario concept   # Relationship decorrelation
prophet drift-demo --scenario regime    # Operational mode transitions

┌─────────────────────────────────────────────────────────────┐
│                     Drift Engine                             │
│                                                              │
│  ┌────────────┐  ┌──────────────┐  ┌────────────────┐      │
│  │ Covariate  │  │ Performance  │  │    Concept     │      │
│  │ Detector   │  │  Detector    │  │   Detector     │      │
│  │            │  │              │  │                │      │
│  │ KS Test    │  │ ADWIN        │  │ Fisher z       │      │
│  │ Page-Hinkl │  │ Error Track  │  │ Correlation    │      │
│  └─────┬──────┘  └──────┬───────┘  └───────┬────────┘      │
│        │                 │                   │               │
│  ┌─────┴─────────────────┴───────────────────┴──────────┐   │
│  │              Regime Detector                          │   │
│  │   Classifies: normal/peak/incident/recovery/...      │   │
│  └─────┬────────────────────────────────────────────────┘   │
│        │                                                     │
│  ┌─────┴────────────────────────────────────────────────┐   │
│  │              Trust Scorer                             │   │
│  │   Per-metric trust (0-100%) from all signals         │   │
│  └─────┬────────────────────────────────────────────────┘   │
│        │                                                     │
│  ┌─────┴────────────────────────────────────────────────┐   │
│  │              Drift Advisor                            │   │
│  │   OK → Monitor → Recalibrate → Retrain → Suspend    │   │
│  └──────────────────────────────────────────────────────┘   │
│                                                              │
│  Persistence: drift.db (SQLite)                             │
└─────────────────────────────────────────────────────────────┘

┌─────────────────────────────────────────────────────────────┐
│                   Fleet Intelligence                         │
│                                                              │
│  ┌──────────────┐  ┌────────────────┐  ┌────────────────┐  │
│  │  Service      │  │  Correlation   │  │  Blast Radius  │  │
│  │  Registry     │  │  Engine        │  │  Estimator     │  │
│  │              │  │                │  │                │  │
│  │  Register    │  │  Cross-corr    │  │  Graph walk    │  │
│  │  Track       │  │  Lag detect    │  │  Impact score  │  │
│  │  Status      │  │  Auto-discover │  │  Cascade sim   │  │
│  └──────┬───────┘  └───────┬────────┘  └───────┬────────┘  │
│         │                  │                     │           │
│  ┌──────┴──────────────────┴─────────────────────┴───────┐  │
│  │              Dependency Graph                          │  │
│  │   Directed graph of service relationships             │  │
│  └──────┬────────────────────────────────────────────────┘  │
│         │                                                    │
│  ┌──────┴────────────────────────────────────────────────┐  │
│  │              Fleet Dashboard                           │  │
│  │   Overall health │ Risk matrix │ Topology view        │  │
│  └──────────────────────────────────────────────────────┘  │
│                                                              │
│  Persistence: fleet.db (SQLite)                             │
└─────────────────────────────────────────────────────────────┘

# Register services with tiers (1=critical, 2=important, 3=normal)
prophet fleet register api-gateway --tier 1 --desc "Main entry point"
prophet fleet register auth-service --tier 1 --desc "Authentication"
prophet fleet register user-service --tier 2 --desc "User management"
prophet fleet register payment-svc --tier 1 --desc "Payment processing"
prophet fleet register notification-svc --tier 3 --desc "Email/SMS"

# Manual: api-gateway depends on auth-service
prophet fleet deps add api-gateway auth-service --type upstream --strength 0.95

# Automatic: discover from metric correlations
prophet fleet discover

prophet fleet status

# Output:
# ══════════════════════════════════════════════════
#   🛰️  FLEET INTELLIGENCE
# ══════════════════════════════════════════════════
#
#   Fleet Health: [████████████████░░░░] 83% (B)
#
#   Services: 7 total
#     🟢 Healthy:  4
#     🟡 Degraded: 1
#     🔴 Critical: 0
#     ❓ Unknown:  2
#
#   ⚠️  Top Risks:
#     🔴 auth-service — widespread (4 services, 82% impact)

prophet fleet blast auth-service

# Output:
# 💥 Blast Radius: auth-service
# ────────────────────────────────────────
#   Severity: 🔴 WIDESPREAD
#   Affected: 4 service(s)
#   Impact:   82%
#   Depth:    2 hop(s)
#
#   📊 Affected Services:
#     api-gateway: [████████████████████] 100% (+0m, depth 1)
#     user-service: [████████████████████] 100% (+0m, depth 1)
#     search-service: [████████████████░░░░] 81% (+0m, depth 2)
#
#   🌊 Cascade Path:
#     auth-service → api-gateway → user-service → search-service
#
#   💡 Mitigations:
#     🚨 Tier-1 critical service — activate incident response
#     ⚡ Critical services at risk: api-gateway — prepare failover
#     🔌 Enable circuit breakers for: api-gateway, user-service

prophet fleet discover --min 0.6

# Output:
# ✅ Discovered 5 cross-service correlations:
#
#   api-gateway.response_time
#     → user-service.response_time
#       [████████░░] 82% (lag: 2.5m, n=150)
#
#   auth-service.cpu
#     → api-gateway.error_rate
#       [███████░░░] 71% (lag: 1.0m, n=200)

# ASCII topology
prophet fleet topology

# Mermaid diagram (for docs/dashboards)
prophet fleet topology --mermaid

prophet fleet risk

# Output:
# 🎯 Risk Matrix
# ────────────────────────────────────────────────
#   Service            Tier  Status    Risk   Blast      Affected
#   user-service        T2    degraded  █████░  widespread 3
#   auth-service        T1    healthy   █░░░░░  widespread 4
#   payment-service     T1    healthy   █░░░░░  moderate   2

# Record events
prophet fleet event deployment api-gateway "Deployed v2.3.1"
prophet fleet event incident auth-service "Elevated latency"

# View recent events
prophet fleet event

prophet fleet compare api-gateway user-service response_time

# Returns: mean, std, p50, p95, p99, correlation, lag

prophet fleet status              # Fleet dashboard
prophet fleet status --json       # JSON output
prophet fleet register <name>     # Register a service
prophet fleet remove <name>       # Remove a service
prophet fleet blast <name>        # Blast radius estimation
prophet fleet discover            # Auto-discover correlations
prophet fleet discover --min 0.7  # Higher correlation threshold
prophet fleet topology            # ASCII topology
prophet fleet topology --mermaid  # Mermaid diagram
prophet fleet risk                # Risk matrix
prophet fleet deps                # All dependencies
prophet fleet deps <name>         # Dependencies for a service
prophet fleet compare <a> <b> <m> # Compare metric across services
prophet fleet event               # Recent events
prophet fleet event <type> <svc> <desc>  # Record event