d14150da75
Implement intelligent agent learning from Knowledge Graph execution history with per-task-type expertise tracking, recency bias, and learning curves. ## Phase 5.3 Implementation ### Learning Infrastructure (✅ Complete) - LearningProfileService with per-task-type expertise metrics - TaskTypeExpertise model tracking success_rate, confidence, learning curves - Recency bias weighting: recent 7 days weighted 3x higher (exponential decay) - Confidence scoring prevents overfitting: min(1.0, executions / 20) - Learning curves computed from daily execution windows ### Agent Scoring Service (✅ Complete) - Unified AgentScore combining SwarmCoordinator + learning profiles - Scoring formula: 0.3*base + 0.5*expertise + 0.2*confidence - Rank agents by combined score for intelligent assignment - Support for recency-biased scoring (recent_success_rate) - Methods: rank_agents, select_best, rank_agents_with_recency ### KG Integration (✅ Complete) - KGPersistence::get_executions_for_task_type() - query by agent + task type - KGPersistence::get_agent_executions() - all executions for agent - Coordinator::load_learning_profile_from_kg() - core KG→Learning integration - Coordinator::load_all_learning_profiles() - batch load for multiple agents - Convert PersistedExecution → ExecutionData for learning calculations ### Agent Assignment Integration (✅ Complete) - AgentCoordinator uses learning profiles for task assignment - extract_task_type() infers task type from title/description - assign_task() scores candidates using AgentScoringService - Fallback to load-based selection if no learning data available - Learning profiles stored in coordinator.learning_profiles RwLock ### Profile Adapter Enhancements (✅ Complete) - create_learning_profile() - initialize empty profiles - add_task_type_expertise() - set task-type expertise - update_profile_with_learning() - update swarm profiles from learning ## Files Modified ### vapora-knowledge-graph/src/persistence.rs (+30 lines) - get_executions_for_task_type(agent_id, task_type, limit) - get_agent_executions(agent_id, limit) ### vapora-agents/src/coordinator.rs (+100 lines) - load_learning_profile_from_kg() - core KG integration method - load_all_learning_profiles() - batch loading for agents - assign_task() already uses learning-based scoring via AgentScoringService ### Existing Complete Implementation - vapora-knowledge-graph/src/learning.rs - calculation functions - vapora-agents/src/learning_profile.rs - data structures and expertise - vapora-agents/src/scoring.rs - unified scoring service - vapora-agents/src/profile_adapter.rs - adapter methods ## Tests Passing - learning_profile: 7 tests ✅ - scoring: 5 tests ✅ - profile_adapter: 6 tests ✅ - coordinator: learning-specific tests ✅ ## Data Flow 1. Task arrives → AgentCoordinator::assign_task() 2. Extract task_type from description 3. Query KG for task-type executions (load_learning_profile_from_kg) 4. Calculate expertise with recency bias 5. Score candidates (SwarmCoordinator + learning) 6. Assign to top-scored agent 7. Execution result → KG → Update learning profiles ## Key Design Decisions ✅ Recency bias: 7-day half-life with 3x weight for recent performance ✅ Confidence scoring: min(1.0, total_executions / 20) prevents overfitting ✅ Hierarchical scoring: 30% base load, 50% expertise, 20% confidence ✅ KG query limit: 100 recent executions per task-type for performance ✅ Async loading: load_learning_profile_from_kg supports concurrent loads ## Next: Phase 5.4 - Cost Optimization Ready to implement budget enforcement and cost-aware provider selection.
341 lines
11 KiB
YAML
341 lines
11 KiB
YAML
apiVersion: provisioning.vapora.io/v1
|
|
kind: Workflow
|
|
metadata:
|
|
name: upgrade-vapora
|
|
description: Rolling upgrade of VAPORA services with zero downtime
|
|
spec:
|
|
version: "0.2.0"
|
|
namespace: vapora-system
|
|
timeout: 1800s # 30 minutes max
|
|
|
|
inputs:
|
|
- name: backend_version
|
|
type: string
|
|
required: true
|
|
description: "Target version for backend service (e.g., 0.3.0)"
|
|
- name: frontend_version
|
|
type: string
|
|
required: true
|
|
description: "Target version for frontend service"
|
|
- name: agents_version
|
|
type: string
|
|
required: true
|
|
description: "Target version for agent runtime"
|
|
- name: upgrade_strategy
|
|
type: string
|
|
required: false
|
|
default: "rolling"
|
|
description: "rolling | blue-green | canary"
|
|
- name: skip_tests
|
|
type: boolean
|
|
required: false
|
|
default: false
|
|
description: "Skip smoke tests before upgrade"
|
|
- name: dry_run
|
|
type: boolean
|
|
required: false
|
|
default: false
|
|
description: "Perform dry-run without actual upgrades"
|
|
|
|
phases:
|
|
|
|
# Phase 1: Pre-upgrade checks
|
|
- name: "Pre-Upgrade Validation"
|
|
description: "Verify cluster health and prepare for upgrade"
|
|
retryable: true
|
|
steps:
|
|
- name: "Check cluster health"
|
|
command: "provisioning health-check --cluster"
|
|
timeout: 300s
|
|
|
|
- name: "Backup current state"
|
|
command: |
|
|
provisioning backup create --cluster vapora-cluster \
|
|
--label pre-upgrade-$(date +%Y%m%d-%H%M%S)
|
|
timeout: 600s
|
|
|
|
- name: "Verify all services are running"
|
|
command: "provisioning health-check --services all --strict"
|
|
timeout: 300s
|
|
|
|
- name: "Create git tag for current state"
|
|
command: |
|
|
CURRENT_BACKEND=$(kubectl get deployment vapora-backend -n vapora-system -o jsonpath='{.spec.template.spec.containers[0].image}')
|
|
git tag -a "pre-upgrade-$(echo $CURRENT_BACKEND | cut -d: -f2)" -m "Pre-upgrade checkpoint"
|
|
timeout: 60s
|
|
|
|
# Phase 2: Drain traffic gracefully
|
|
- name: "Prepare for Upgrade"
|
|
description: "Gracefully drain and prepare services for upgrade"
|
|
retryable: true
|
|
steps:
|
|
- name: "Drain agent queue"
|
|
command: |
|
|
provisioning agents drain --timeout 600s \
|
|
--allow-new-work false
|
|
timeout: 700s
|
|
|
|
- name: "Enable maintenance mode"
|
|
command: |
|
|
kubectl patch configmap vapora-config \
|
|
-n vapora-system \
|
|
-p '{"data":{"maintenance_mode":"true"}}'
|
|
timeout: 60s
|
|
|
|
- name: "Wait for in-flight requests to complete"
|
|
command: |
|
|
provisioning metrics wait-for \
|
|
--metric "http_requests_in_flight" \
|
|
--target 0 \
|
|
--timeout 300s
|
|
timeout: 320s
|
|
|
|
# Phase 3: Database migration (if needed)
|
|
- name: "Database Migrations"
|
|
description: "Apply database schema changes"
|
|
retryable: false
|
|
steps:
|
|
- name: "Create database backup"
|
|
command: |
|
|
provisioning db backup --database surrealdb \
|
|
--output backup-pre-upgrade-$(date +%s).sql
|
|
timeout: 600s
|
|
|
|
- name: "Run migration scripts"
|
|
command: |
|
|
for MIGRATION in scripts/migrations/v0.3.0/*.surql; do
|
|
echo "Running migration: $MIGRATION"
|
|
provisioning db execute --database surrealdb --file "$MIGRATION" || {
|
|
echo "Migration failed, restoring backup"
|
|
exit 1
|
|
}
|
|
done
|
|
timeout: 600s
|
|
|
|
- name: "Verify migration success"
|
|
command: "provisioning db verify --database surrealdb"
|
|
timeout: 300s
|
|
|
|
# Phase 4: Update backend service
|
|
- name: "Upgrade Backend Service"
|
|
description: "Rolling update of REST API backend"
|
|
retryable: true
|
|
steps:
|
|
- name: "Update backend image"
|
|
command: |
|
|
if [ "$DRY_RUN" = "true" ]; then
|
|
echo "[DRY-RUN] Would update backend to vapora/backend:$BACKEND_VERSION"
|
|
else
|
|
provisioning taskserv upgrade vapora-backend \
|
|
--image vapora/backend:$BACKEND_VERSION \
|
|
--strategy rolling \
|
|
--max-surge 1 \
|
|
--max-unavailable 0
|
|
fi
|
|
timeout: 600s
|
|
env:
|
|
- name: BACKEND_VERSION
|
|
value: "${backend_version}"
|
|
- name: DRY_RUN
|
|
value: "${dry_run}"
|
|
|
|
- name: "Wait for backend rollout"
|
|
command: "kubectl rollout status deployment/vapora-backend -n vapora-system --timeout=300s"
|
|
timeout: 320s
|
|
|
|
- name: "Run smoke tests"
|
|
command: |
|
|
if [ "$SKIP_TESTS" != "true" ]; then
|
|
provisioning test smoke --api http://vapora-backend.vapora-system:8080 \
|
|
--endpoints "/api/v1/health" "/api/v1/ready"
|
|
fi
|
|
timeout: 180s
|
|
env:
|
|
- name: SKIP_TESTS
|
|
value: "${skip_tests}"
|
|
continueOnError: true
|
|
|
|
# Phase 5: Update LLM Router and MCP Gateway
|
|
- name: "Upgrade Backend Components"
|
|
description: "Update LLM Router and MCP Gateway in parallel"
|
|
retryable: true
|
|
parallel: true
|
|
steps:
|
|
- name: "Upgrade LLM Router"
|
|
command: |
|
|
if [ "$DRY_RUN" != "true" ]; then
|
|
provisioning taskserv upgrade vapora-llm-router \
|
|
--image vapora/llm-router:$VERSION \
|
|
--strategy rolling \
|
|
--max-unavailable 0
|
|
fi
|
|
timeout: 600s
|
|
env:
|
|
- name: VERSION
|
|
value: "${backend_version}"
|
|
|
|
- name: "Upgrade MCP Gateway"
|
|
command: |
|
|
if [ "$DRY_RUN" != "true" ]; then
|
|
provisioning taskserv upgrade vapora-mcp-gateway \
|
|
--image vapora/mcp-gateway:$VERSION \
|
|
--strategy rolling \
|
|
--max-unavailable 0
|
|
fi
|
|
timeout: 600s
|
|
env:
|
|
- name: VERSION
|
|
value: "${backend_version}"
|
|
|
|
# Phase 6: Update agent runtime
|
|
- name: "Upgrade Agent Runtime"
|
|
description: "Update agent runtime with safe rollout"
|
|
retryable: true
|
|
steps:
|
|
- name: "Update agent image"
|
|
command: |
|
|
if [ "$DRY_RUN" != "true" ]; then
|
|
provisioning taskserv upgrade vapora-agents \
|
|
--image vapora/agents:$VERSION \
|
|
--strategy rolling \
|
|
--max-surge 1 \
|
|
--max-unavailable 1 \
|
|
--drain-timeout 300s
|
|
fi
|
|
timeout: 900s
|
|
env:
|
|
- name: VERSION
|
|
value: "${agents_version}"
|
|
|
|
- name: "Wait for agents to stabilize"
|
|
command: |
|
|
kubectl wait --for=condition=Ready pod \
|
|
-l app=vapora-agents \
|
|
-n vapora-system \
|
|
--timeout=600s
|
|
timeout: 620s
|
|
|
|
# Phase 7: Update frontend service
|
|
- name: "Upgrade Frontend Service"
|
|
description: "Update UI frontend with minimal user impact"
|
|
retryable: true
|
|
steps:
|
|
- name: "Update frontend image"
|
|
command: |
|
|
if [ "$DRY_RUN" != "true" ]; then
|
|
provisioning taskserv upgrade vapora-frontend \
|
|
--image vapora/frontend:$VERSION \
|
|
--strategy rolling \
|
|
--max-surge 1 \
|
|
--max-unavailable 0
|
|
fi
|
|
timeout: 600s
|
|
env:
|
|
- name: VERSION
|
|
value: "${frontend_version}"
|
|
|
|
- name: "Wait for frontend rollout"
|
|
command: "kubectl rollout status deployment/vapora-frontend -n vapora-system --timeout=300s"
|
|
timeout: 320s
|
|
|
|
- name: "Test frontend endpoints"
|
|
command: |
|
|
if [ "$SKIP_TESTS" != "true" ]; then
|
|
provisioning test smoke --frontend http://vapora-frontend.vapora-system:3000 \
|
|
--endpoints "/"
|
|
fi
|
|
timeout: 180s
|
|
|
|
# Phase 8: Post-upgrade verification
|
|
- name: "Post-Upgrade Verification"
|
|
description: "Comprehensive validation of upgraded system"
|
|
retryable: false
|
|
steps:
|
|
- name: "Disable maintenance mode"
|
|
command: |
|
|
kubectl patch configmap vapora-config \
|
|
-n vapora-system \
|
|
-p '{"data":{"maintenance_mode":"false"}}'
|
|
timeout: 60s
|
|
|
|
- name: "Health check all services"
|
|
command: "provisioning health-check --services all --strict"
|
|
timeout: 300s
|
|
|
|
- name: "Verify agent communication"
|
|
command: "provisioning agents health-check --nats nats://nats-0.vapora-system:4222"
|
|
timeout: 120s
|
|
|
|
- name: "Run integration tests"
|
|
command: "provisioning test integration --timeout 600s"
|
|
timeout: 620s
|
|
continueOnError: true
|
|
|
|
- name: "Check application logs for errors"
|
|
command: |
|
|
ERROR_COUNT=$(kubectl logs -n vapora-system -l app=vapora-backend --tail=1000 | grep -c 'ERROR\|CRITICAL')
|
|
if [ "$ERROR_COUNT" -gt 10 ]; then
|
|
echo "WARNING: Found $ERROR_COUNT errors in backend logs"
|
|
exit 1
|
|
fi
|
|
timeout: 120s
|
|
continueOnError: true
|
|
|
|
- name: "Re-enable agent work"
|
|
command: "provisioning agents drain --disable"
|
|
timeout: 60s
|
|
|
|
# Phase 9: Tag and document upgrade
|
|
- name: "Finalize Upgrade"
|
|
description: "Document upgrade completion"
|
|
retryable: false
|
|
steps:
|
|
- name: "Create upgrade completion tag"
|
|
command: |
|
|
git tag -a "upgraded-to-$BACKEND_VERSION-$(date +%Y%m%d-%H%M%S)" \
|
|
-m "Upgrade completed: backend=$BACKEND_VERSION, frontend=$FRONTEND_VERSION, agents=$AGENTS_VERSION"
|
|
timeout: 60s
|
|
env:
|
|
- name: BACKEND_VERSION
|
|
value: "${backend_version}"
|
|
- name: FRONTEND_VERSION
|
|
value: "${frontend_version}"
|
|
- name: AGENTS_VERSION
|
|
value: "${agents_version}"
|
|
|
|
- name: "Generate upgrade report"
|
|
command: |
|
|
provisioning report generate \
|
|
--type upgrade \
|
|
--format markdown \
|
|
--output "upgrade-report-$(date +%Y%m%d-%H%M%S).md"
|
|
timeout: 120s
|
|
|
|
# Rollback procedure
|
|
onFailure:
|
|
rollback: true
|
|
procedure:
|
|
- name: "Restore from pre-upgrade backup"
|
|
command: "provisioning backup restore --label pre-upgrade-* --latest"
|
|
- name: "Verify rollback success"
|
|
command: "provisioning health-check --cluster"
|
|
|
|
outputs:
|
|
- name: upgrade_status
|
|
value: "echo 'Upgrade completed successfully'"
|
|
- name: versions_deployed
|
|
command: "kubectl get deployment -n vapora-system -o wide"
|
|
|
|
notifications:
|
|
onStart:
|
|
- "slack: #deployment"
|
|
- "email: devops@example.com"
|
|
onSuccess:
|
|
- "slack: #deployment"
|
|
- "slack: notify: Upgrade successful"
|
|
onFailure:
|
|
- "slack: #deployment"
|
|
- "slack: #alerts"
|
|
- "email: devops@example.com"
|
|
- "severity: critical"
|