agentic-qe 3.7.8 → 3.7.10

package/assets/governance/shards/chaos-resilience.shard.md

@@ -1,221 +1,221 @@
-# Chaos Resilience Domain Shard
-
-**Domain**: chaos-resilience
-**Version**: 1.0.0
-**Last Updated**: 2026-02-03
-**Parent Constitution**: `.claude/guidance/constitution.md`
-
----
-
-## Domain Rules
-
-1. **Hypothesis-Driven Experiments**: All chaos experiments MUST have a defined hypothesis with measurable steady-state indicators; exploratory chaos is prohibited in production.
-
-2. **Blast Radius Control**: Chaos experiments MUST define and enforce blast radius limits; unconstrained fault injection is a constitutional violation.
-
-3. **Rollback Plan Required**: Every chaos experiment MUST have an automated rollback plan with defined triggers; experiments without rollback are rejected.
-
-4. **Production Gating**: Chaos experiments in production MUST pass staging validation first; direct-to-production experiments require explicit approval.
-
-5. **Load Test Baseline**: Performance load tests MUST establish baselines before comparative analysis; baseline-less load tests produce invalid conclusions.
-
-6. **Circuit Breaker Verification**: Resilience testing MUST verify circuit breaker behavior under fault conditions; untested circuit breakers are deployment blockers.
-
----
-
-## Quality Thresholds
-
-| Metric | Minimum | Target | Critical |
-|--------|---------|--------|----------|
-| Experiment Success Rate | 0.8 | 0.95 | < 0.6 |
-| Recovery Time (MTTR) | < 5 min | < 2 min | > 15 min |
-| Blast Radius Control | 100% | 100% | < 90% |
-| Load Test Reliability | 0.9 | 0.98 | < 0.8 |
-| Circuit Breaker Coverage | 0.8 | 0.95 | < 0.6 |
-| Hypothesis Validation | 0.7 | 0.85 | < 0.5 |
-
----
-
-## Invariants
-
-```
-INVARIANT hypothesis_required:
-  FOR ALL experiment IN chaos_experiments:
-    experiment.hypothesis IS NOT NULL AND
-    experiment.steady_state_definition IS NOT NULL AND
-    experiment.steady_state_definition.probes.length >= 1
-```
-
-```
-INVARIANT blast_radius_control:
-  FOR ALL experiment IN chaos_experiments:
-    experiment.blast_radius IS NOT NULL AND
-    experiment.blast_radius.affected_services.length <= experiment.blast_radius.max_services AND
-    experiment.blast_radius.affected_users_percent <= experiment.blast_radius.max_users_percent
-```
-
-```
-INVARIANT rollback_plan_required:
-  FOR ALL experiment IN chaos_experiments:
-    experiment.rollback_plan IS NOT NULL AND
-    experiment.rollback_plan.triggers.length >= 1 AND
-    experiment.rollback_plan.automated = true
-```
-
-```
-INVARIANT staging_before_production:
-  FOR ALL prod_experiment IN production_experiments:
-    EXISTS staging_run WHERE
-      staging_run.experiment_type = prod_experiment.type AND
-      staging_run.passed = true AND
-      staging_run.timestamp < prod_experiment.timestamp
-```
-
-```
-INVARIANT load_test_baseline:
-  FOR ALL load_test IN comparative_load_tests:
-    load_test.baseline_established = true AND
-    load_test.baseline_timestamp IS NOT NULL
-```
-
----
-
-## Patterns
-
-**Domain Source**: `v3/src/domains/chaos-resilience/`
-
-| Pattern | Location | Description |
-|---------|----------|-------------|
-| Chaos Engineer Service | `services/chaos-engineer.ts` | Fault injection orchestration |
-| Load Tester Service | `services/load-tester.ts` | Performance load testing |
-| Performance Profiler Service | `services/performance-profiler.ts` | Bottleneck detection |
-| Chaos Resilience Coordinator | `coordinator.ts` | Workflow orchestration |
-
-**Key Interfaces**: `interfaces/index.ts` defines `ChaosExperiment`, `LoadTest`, `BlastRadius`, `RollbackPlan`, and related types.
-
----
-
-## Agent Constraints
-
-| Role | Agent ID | Permissions |
-|------|----------|-------------|
-| **Primary** | `qe-chaos-engineer` | Full experiment design and execution |
-| **Secondary** | `qe-performance-tester` | Load testing, profiling |
-| **Secondary** | `qe-resilience-assessor` | Recovery testing |
-| **Support** | `qe-code-analyst` | Service dependency mapping |
-| **Approval** | `sre-team` (human) | Production chaos approval |
-
-**Forbidden Actions**: No agent may execute production chaos experiments without staging validation and explicit approval.
-
----
-
-## Escalation Triggers
-
-| Trigger | Severity | Action |
-|---------|----------|--------|
-| Blast radius exceeded | CRITICAL | Immediate rollback, escalate to SRE team |
-| MTTR > 15 min | CRITICAL | Halt experiment, escalate to Queen Coordinator |
-| Rollback failure | CRITICAL | Manual intervention required, escalate immediately |
-| Production chaos without staging | CRITICAL | Block execution, escalate |
-| Hypothesis invalidated | HIGH | Escalate for experiment redesign |
-| Circuit breaker failure | HIGH | Escalate, flag service for remediation |
-| Load test baseline missing | MEDIUM | Block comparative analysis |
-| Experiment success rate < 0.6 | HIGH | Pause experiments, review design |
-
----
-
-## Memory Namespace
-
-- **Namespace**: `qe-patterns/chaos-resilience`
-- **Retention**: 90 days (experiment results retained longer)
-- **Contradiction Check**: Enabled
-
----
-
-## Integration Points
-
-| Domain | Integration Type | Purpose |
-|--------|-----------------|---------|
-| `code-intelligence` | Input | Service dependency graphs |
-| `quality-assessment` | Output | Report resilience scores |
-| `test-execution` | Bidirectional | Execute resilience tests |
-| `defect-intelligence` | Output | Report failure patterns |
-| `learning-optimization` | Bidirectional | Share resilience patterns |
-
----
-
-## Fault Types
-
-| Fault Type | Description | Blast Radius Risk |
-|------------|-------------|-------------------|
-| network-latency | Add latency to network calls | Medium |
-| network-partition | Isolate services | High |
-| cpu-stress | Exhaust CPU resources | Medium |
-| memory-pressure | Consume memory | Medium |
-| disk-io | Slow disk operations | Low |
-| process-kill | Terminate processes | High |
-| dns-failure | Break DNS resolution | High |
-| clock-skew | Advance/retard system clock | Low |
-
----
-
-## Load Test Types
-
-| Type | Description | Use Case |
-|------|-------------|----------|
-| smoke | Basic functionality under load | Pre-deployment validation |
-| load | Normal expected load | Baseline establishment |
-| stress | Beyond normal load | Find breaking points |
-| spike | Sudden traffic increase | Autoscaling validation |
-| soak | Extended duration load | Memory leak detection |
-
----
-
-## Rollback Trigger Schema
-
-```typescript
-interface RollbackTrigger {
-  metric: string;
-  operator: 'gt' | 'lt' | 'eq' | 'gte' | 'lte';
-  threshold: number;
-  duration: string;
-  action: 'rollback' | 'pause' | 'alert';
-  automatic: boolean;
-}
-```
-
----
-
-## Blast Radius Definition
-
-```typescript
-interface BlastRadius {
-  maxServices: number;
-  maxUsersPercent: number;
-  excludedServices: string[];
-  geographicScope: 'single-region' | 'multi-region' | 'global';
-  timeWindow: {
-    start: string;
-    end: string;
-    timezone: string;
-  };
-  trafficPercentage: number;
-}
-```
-
----
-
-## GameDay Protocol
-
-1. **Planning**: Define hypothesis, metrics, rollback criteria
-2. **Communication**: Notify all stakeholders 24h before
-3. **Staging Run**: Execute in staging environment
-4. **Review**: Analyze staging results
-5. **Production Run**: Execute with blast radius controls
-6. **Monitoring**: Real-time observation dashboard
-7. **Rollback**: If triggers activated, automatic rollback
-8. **Post-mortem**: Document findings, update resilience patterns
-
----
-
-*This shard is enforced by @claude-flow/guidance governance system.*
+# Chaos Resilience Domain Shard
+
+**Domain**: chaos-resilience
+**Version**: 1.0.0
+**Last Updated**: 2026-02-03
+**Parent Constitution**: `.claude/guidance/constitution.md`
+
+---
+
+## Domain Rules
+
+1. **Hypothesis-Driven Experiments**: All chaos experiments MUST have a defined hypothesis with measurable steady-state indicators; exploratory chaos is prohibited in production.
+
+2. **Blast Radius Control**: Chaos experiments MUST define and enforce blast radius limits; unconstrained fault injection is a constitutional violation.
+
+3. **Rollback Plan Required**: Every chaos experiment MUST have an automated rollback plan with defined triggers; experiments without rollback are rejected.
+
+4. **Production Gating**: Chaos experiments in production MUST pass staging validation first; direct-to-production experiments require explicit approval.
+
+5. **Load Test Baseline**: Performance load tests MUST establish baselines before comparative analysis; baseline-less load tests produce invalid conclusions.
+
+6. **Circuit Breaker Verification**: Resilience testing MUST verify circuit breaker behavior under fault conditions; untested circuit breakers are deployment blockers.
+
+---
+
+## Quality Thresholds
+
+| Metric | Minimum | Target | Critical |
+|--------|---------|--------|----------|
+| Experiment Success Rate | 0.8 | 0.95 | < 0.6 |
+| Recovery Time (MTTR) | < 5 min | < 2 min | > 15 min |
+| Blast Radius Control | 100% | 100% | < 90% |
+| Load Test Reliability | 0.9 | 0.98 | < 0.8 |
+| Circuit Breaker Coverage | 0.8 | 0.95 | < 0.6 |
+| Hypothesis Validation | 0.7 | 0.85 | < 0.5 |
+
+---
+
+## Invariants
+
+```
+INVARIANT hypothesis_required:
+  FOR ALL experiment IN chaos_experiments:
+    experiment.hypothesis IS NOT NULL AND
+    experiment.steady_state_definition IS NOT NULL AND
+    experiment.steady_state_definition.probes.length >= 1
+```
+
+```
+INVARIANT blast_radius_control:
+  FOR ALL experiment IN chaos_experiments:
+    experiment.blast_radius IS NOT NULL AND
+    experiment.blast_radius.affected_services.length <= experiment.blast_radius.max_services AND
+    experiment.blast_radius.affected_users_percent <= experiment.blast_radius.max_users_percent
+```
+
+```
+INVARIANT rollback_plan_required:
+  FOR ALL experiment IN chaos_experiments:
+    experiment.rollback_plan IS NOT NULL AND
+    experiment.rollback_plan.triggers.length >= 1 AND
+    experiment.rollback_plan.automated = true
+```
+
+```
+INVARIANT staging_before_production:
+  FOR ALL prod_experiment IN production_experiments:
+    EXISTS staging_run WHERE
+      staging_run.experiment_type = prod_experiment.type AND
+      staging_run.passed = true AND
+      staging_run.timestamp < prod_experiment.timestamp
+```
+
+```
+INVARIANT load_test_baseline:
+  FOR ALL load_test IN comparative_load_tests:
+    load_test.baseline_established = true AND
+    load_test.baseline_timestamp IS NOT NULL
+```
+
+---
+
+## Patterns
+
+**Domain Source**: `v3/src/domains/chaos-resilience/`
+
+| Pattern | Location | Description |
+|---------|----------|-------------|
+| Chaos Engineer Service | `services/chaos-engineer.ts` | Fault injection orchestration |
+| Load Tester Service | `services/load-tester.ts` | Performance load testing |
+| Performance Profiler Service | `services/performance-profiler.ts` | Bottleneck detection |
+| Chaos Resilience Coordinator | `coordinator.ts` | Workflow orchestration |
+
+**Key Interfaces**: `interfaces/index.ts` defines `ChaosExperiment`, `LoadTest`, `BlastRadius`, `RollbackPlan`, and related types.
+
+---
+
+## Agent Constraints
+
+| Role | Agent ID | Permissions |
+|------|----------|-------------|
+| **Primary** | `qe-chaos-engineer` | Full experiment design and execution |
+| **Secondary** | `qe-performance-tester` | Load testing, profiling |
+| **Secondary** | `qe-resilience-assessor` | Recovery testing |
+| **Support** | `qe-code-analyst` | Service dependency mapping |
+| **Approval** | `sre-team` (human) | Production chaos approval |
+
+**Forbidden Actions**: No agent may execute production chaos experiments without staging validation and explicit approval.
+
+---
+
+## Escalation Triggers
+
+| Trigger | Severity | Action |
+|---------|----------|--------|
+| Blast radius exceeded | CRITICAL | Immediate rollback, escalate to SRE team |
+| MTTR > 15 min | CRITICAL | Halt experiment, escalate to Queen Coordinator |
+| Rollback failure | CRITICAL | Manual intervention required, escalate immediately |
+| Production chaos without staging | CRITICAL | Block execution, escalate |
+| Hypothesis invalidated | HIGH | Escalate for experiment redesign |
+| Circuit breaker failure | HIGH | Escalate, flag service for remediation |
+| Load test baseline missing | MEDIUM | Block comparative analysis |
+| Experiment success rate < 0.6 | HIGH | Pause experiments, review design |
+
+---
+
+## Memory Namespace
+
+- **Namespace**: `qe-patterns/chaos-resilience`
+- **Retention**: 90 days (experiment results retained longer)
+- **Contradiction Check**: Enabled
+
+---
+
+## Integration Points
+
+| Domain | Integration Type | Purpose |
+|--------|-----------------|---------|
+| `code-intelligence` | Input | Service dependency graphs |
+| `quality-assessment` | Output | Report resilience scores |
+| `test-execution` | Bidirectional | Execute resilience tests |
+| `defect-intelligence` | Output | Report failure patterns |
+| `learning-optimization` | Bidirectional | Share resilience patterns |
+
+---
+
+## Fault Types
+
+| Fault Type | Description | Blast Radius Risk |
+|------------|-------------|-------------------|
+| network-latency | Add latency to network calls | Medium |
+| network-partition | Isolate services | High |
+| cpu-stress | Exhaust CPU resources | Medium |
+| memory-pressure | Consume memory | Medium |
+| disk-io | Slow disk operations | Low |
+| process-kill | Terminate processes | High |
+| dns-failure | Break DNS resolution | High |
+| clock-skew | Advance/retard system clock | Low |
+
+---
+
+## Load Test Types
+
+| Type | Description | Use Case |
+|------|-------------|----------|
+| smoke | Basic functionality under load | Pre-deployment validation |
+| load | Normal expected load | Baseline establishment |
+| stress | Beyond normal load | Find breaking points |
+| spike | Sudden traffic increase | Autoscaling validation |
+| soak | Extended duration load | Memory leak detection |
+
+---
+
+## Rollback Trigger Schema
+
+```typescript
+interface RollbackTrigger {
+  metric: string;
+  operator: 'gt' | 'lt' | 'eq' | 'gte' | 'lte';
+  threshold: number;
+  duration: string;
+  action: 'rollback' | 'pause' | 'alert';
+  automatic: boolean;
+}
+```
+
+---
+
+## Blast Radius Definition
+
+```typescript
+interface BlastRadius {
+  maxServices: number;
+  maxUsersPercent: number;
+  excludedServices: string[];
+  geographicScope: 'single-region' | 'multi-region' | 'global';
+  timeWindow: {
+    start: string;
+    end: string;
+    timezone: string;
+  };
+  trafficPercentage: number;
+}
+```
+
+---
+
+## GameDay Protocol
+
+1. **Planning**: Define hypothesis, metrics, rollback criteria
+2. **Communication**: Notify all stakeholders 24h before
+3. **Staging Run**: Execute in staging environment
+4. **Review**: Analyze staging results
+5. **Production Run**: Execute with blast radius controls
+6. **Monitoring**: Real-time observation dashboard
+7. **Rollback**: If triggers activated, automatic rollback
+8. **Post-mortem**: Document findings, update resilience patterns
+
+---
+
+*This shard is enforced by @claude-flow/guidance governance system.*