Real-World Integration Patterns & Deployment

Common Integration Patterns

Pattern 1: Salesforce ↔ Genesys Contact Sync

Scenario: Synchronize Salesforce contacts to Genesys external contacts

Architecture:

Genesys Cloud API
       ↑
       │ PATCH /externalcontacts
       │ (partial updates, Mar 2026)
       │
   Sync Service
   (Node.js/Python)
       ↑
       │ Query Salesforce API
       │
Salesforce Org

Frequency: Every 30 minutes
Direction: Salesforce → Genesys (one-way)
Trigger: Scheduled job (cron)
Volume: ~1,000 contacts per sync

Authentication:

Salesforce:
├─ OAuth 2.0 (your existing setup)
├─ Service account or user credentials
└─ Connected app registered

Genesys:
├─ OAuth 2.0 Client Credentials
├─ Role: External Contact Manager
├─ Scopes: externalcontacts:manage
└─ Monthly secret rotation

Data Flow:

1. Query Salesforce:
   GET /services/data/v60.0/query
   SELECT Id, Email, Phone, Name, AccountId
   FROM Contact
   WHERE LastModifiedDate > :lastSync

2. Transform Data:
   ├─ Normalize phone (E.164)
   ├─ Validate email
   ├─ Map custom fields
   ├─ Add external ID (Salesforce ID)
   └─ Group by action (create/update)

3. Sync to Genesys:
   ├─ New contacts: POST /externalcontacts/contacts
   ├─ Updates: PATCH /externalcontacts/contacts/{id}
   ├─ Batches: Group every 50-100
   └─ Handle errors: Log and retry

4. Track Progress:
   ├─ Timestamp last successful sync
   ├─ Count created/updated/failed
   ├─ Send email notification
   └─ Log all activities

Implementation Highlights:

Error Handling:
├─ 409 Conflict: Already exists (update instead)
├─ 400 Bad Request: Invalid data (log and skip)
├─ 429 Rate Limited: Backoff and retry
├─ 503 Unavailable: Retry with backoff
└─ All other: Fail and alert

Partial Updates (PATCH):
├─ Update only changed fields
├─ Prevents overwriting unmanaged data
├─ Reduces payload size
├─ Better for CRM sync
└─ Available March 2026+

Idempotency:
├─ Use externalId for matching
├─ Prevent duplicate creates
├─ Retry-safe operations
├─ Track processed records
└─ Handle partial failures

Performance:
├─ Batch 50 contacts per request
├─ 1,000 contacts: ~20 requests
├─ Completes in < 5 minutes
├─ Well under rate limits
└─ Minimal API footprint

Benefits:
├─ Single source of truth
├─ Real-time contact availability
├─ No duplicate entry
├─ Reduces manual effort
└─ Improved agent experience

Pattern 2: Nightly Analytics Report Generation

Scenario: Generate daily contact center reports

Architecture:

Genesys Cloud Analytics API
         ↑
         │
Report Generator Service
(scheduled, 2:00 AM daily)
         ↓
         │
Email Distribution List

Authentication:

Client Credentials:
├─ Scopes: analytics:conversationDetail
├─ Token Duration: 1 hour (sufficient)
├─ No user interaction needed
└─ Fully automated

Workflow:

1. Authenticate (2:00 AM)
   POST /oauth/token
   grant_type: client_credentials
   Result: Access token

2. Query Analytics (2:01 AM)
   ├─ Yesterday's date range
   ├─ All queues/teams
   ├─ Conversations aggregate
   ├─ Service Level, AHT, ASA, Abandon Rate
   └─ Multiple requests (by dimension)

3. Generate Report (2:05 AM)
   ├─ Process data
   ├─ Create charts/graphs
   ├─ Format professionally
   ├─ Create PDF
   └─ Calculate YoY trends

4. Distribute (2:10 AM)
   ├─ Email PDF
   ├─ To stakeholders
   ├─ With summary
   └─ Archive for history

Report Contents:

Executive Summary:
├─ Total conversations: 5,432
├─ Service Level: 87% (Target: 80%)
├─ Avg Handle Time: 8:32 (Target: < 10min)
├─ Abandon Rate: 3.2% (Target: < 5%)
└─ Net Change vs yesterday: +2.1%

By Queue:
├─ Queue name
├─ Conversations
├─ Service Level
├─ AHT
├─ ASA
└─ Agents

Trending:
├─ Last 7 days
├─ Last 30 days
├─ YoY comparison
├─ Alerts (SLA failures)
└─ Recommendations

Implementation Highlights:

Scheduling:
├─ Cron: "0 2 * * *" (2:00 AM daily)
├─ Timezone: Your organization's TZ
├─ Alerting: If job fails
└─ Retry: Automatic

API Calls:
├─ /analytics/conversations/aggregates (multi-call)
├─ Total: 5-10 requests
├─ Completes in < 5 minutes
└─ No rate limit issues

Report Generation:
├─ Tool: ReportLab (Python) or similar
├─ Format: PDF
├─ Design: Professional template
└─ Data: Charts and tables

Email Distribution:
├─ Tool: SendGrid or SMTP
├─ Recipients: DL
├─ Schedule: Exactly 2:15 AM
├─ Archive: Save copy for history
└─ Tracking: Note delivery status

Error Handling:
├─ Authentication fails: Alert operations
├─ API call fails: Retry with backoff
├─ Report gen fails: Fallback to text
├─ Email fails: Retry next hour
└─ All failures: Log and notify

Benefits:
├─ Automated daily reporting
├─ Saves analyst 30 minutes/day
├─ Consistent delivery
├─ Stakeholders stay informed
└─ Data-driven decisions

Pattern 3: Real-Time Agent Status to Dashboard

Scenario: Display real-time agent availability in web dashboard

Architecture:

Genesys Cloud (WebSocket)
       ↓ /v2/users/{id}/presence
       │ (WebSocket events)
       ↓
Node.js Backend
(WebSocket server)
       ↓ Socket.IO
       ↓
Browser Clients
(Dashboard)

Authentication:

User Login:
├─ Authorization Code Grant
├─ User authenticates once
├─ Backend stores refresh_token
├─ Token includes necessary scopes
└─ presence:manage scope required

WebSocket Connection:
├─ Uses existing access_token
├─ Maintains connection
├─ Real-time events
└─ Automatic reconnection

Workflow:

1. User Logs In to Dashboard:
   ├─ Authorization Code flow
   ├─ User sees consent screen
   ├─ Backend gets access_token
   ├─ Backend stores refresh_token
   └─ User authenticated

2. Backend Establishes WebSocket:
   POST /api/v2/notifications/channels
   └─ Opens long-lived connection

3. Subscribe to Presence Events:
   PUT /api/v2/notifications/channels/{channelId}/subscriptions
   
   Subscriptions:
   ├─ v2.users.{user1}.presence
   ├─ v2.users.{user2}.presence
   ├─ v2.users.{user3}.presence
   └─ For all agents

4. Receive Real-Time Events:
   {
     "eventBody": {
       "userId": "user-123",
       "presenceDefinition": {
         "systemPresence": "available",
         "customPresences": [...]
       }
     }
   }

5. Update Dashboard:
   ├─ Forward event to browsers (Socket.IO)
   ├─ Update agent status display
   ├─ Change color/icon
   ├─ Show "Available", "Break", "Busy", etc
   └─ Real-time synchronization

Implementation Highlights:

Subscription Efficiency:
├─ Single WebSocket connection
├─ Multiple subscriptions
├─ vs: 100 polling requests/minute
├─ 99% reduction in API calls
└─ Real-time delivery

Connection Management:
├─ Maintain connection
├─ Automatic reconnection on failure
├─ Health checks
├─ Graceful degradation
└─ Error handling

Scalability:
├─ 1 backend: 100+ agent subscriptions
├─ vs: Polling → rate limited
├─ WebSocket: Event-driven
├─ Lower CPU usage
└─ Lower bandwidth

Error Recovery:
├─ Connection drops: Auto-reconnect
├─ Subscription fails: Retry
├─ Event deserialization: Log and skip
├─ Token expires: Refresh and reconnect
└─ Graceful fallback: Polling backup

Benefits:
├─ Real-time agent status
├─ Sub-second updates
├─ No polling overhead
├─ Better user experience
├─ Reduced infrastructure load
└─ Scalable solution

Deployment Strategies

Strategy 1: Development Environment

Setup:

OAuth Client:
├─ Grant Type: Authorization Code + PKCE
├─ Redirect URI: http://localhost:3000/callback
├─ Scopes: conversations:readonly
├─ Token Duration: 3600 (1 hour)
└─ Created by: Development team

Client Credentials (if needed):
├─ Grant Type: Client Credentials
├─ Scopes: externalcontacts:manage
├─ Token Duration: 3600
└─ Role: External Contact Manager (dev only)

Secrets Storage:
├─ .env file (local development)
├─ .gitignore entries:
│  ├─ .env
│  ├─ .env.local
│  └─ credentials/
├─ Example .env:
│  ├─ GENESYS_CLIENT_ID=dev-client-id
│  ├─ GENESYS_CLIENT_SECRET=dev-secret
│  └─ GENESYS_REGION=mypurecloud.com
└─ Never commit secrets!

Configuration:
├─ Use environment variables
├─ Different per developer
├─ Allow local overrides
├─ Development region specified
└─ Non-production data only

Testing:
├─ Unit tests: Mock API responses
├─ Integration tests: Dev Genesys org
├─ Manual testing: Full flow
├─ Error scenario testing
└─ Rate limit testing

CI/CD Pipeline:
├─ Run on: Developer machine
├─ Linting: Yes
├─ Unit tests: Yes
├─ Build: Yes
├─ Deploy: Local only
└─ Secrets: Via .env (not checked in)

Monitoring:
├─ Logging: Console output
├─ Debugging: Browser DevTools
├─ API calls: Inspect requests
├─ Errors: Stack traces
└─ Performance: Basic measurements

Strategy 2: Staging Environment

Setup:

OAuth Clients:
├─ Separate from production
├─ Authorization Code client
├─ Client Credentials client
├─ Different secrets
└─ Same Genesys region (or test region)

Secrets Storage:
├─ Environment variables (CI/CD platform)
├─ Encrypted vault
├─ GitHub Actions secrets / GitLab CI variables
├─ Rotate monthly
└─ Different from production

Configuration:
├─ Staging-specific settings
├─ Same region as production
├─ Test data only
├─ Verbose logging enabled
└─ Performance testing configured

Testing:
├─ Full integration tests
├─ End-to-end workflows
├─ Load testing (small scale)
├─ Error scenario testing
├─ Performance baselines
└─ Compatibility testing

CI/CD Pipeline:
├─ Trigger: On PR/Merge to staging branch
├─ Linting: Yes
├─ Unit tests: Yes
├─ Integration tests: Yes
├─ Build: Yes
├─ Deploy: Automated
├─ Smoke tests: Post-deploy
└─ Notify: On success/failure

Monitoring:
├─ Application logs: Aggregated
├─ Error tracking: Sentry/similar
├─ Performance monitoring: APM
├─ Uptime monitoring: Synthetic
└─ Alerts: To development team

Approval Process:
├─ Code review: Required
├─ Tests: Must pass
├─ Deployment: Semi-automated
├─ Rollback: Available
└─ Notify stakeholders

Data Management:
├─ Test data only
├─ Reset daily/weekly
├─ Production data: Never
├─ GDPR compliant
└─ Privacy respected

Strategy 3: Production Environment

Setup:

OAuth Clients:
├─ Separate production clients
├─ Multiple clients: Web app, services, etc.
├─ Different secrets per environment
├─ Rotate monthly minimum
├─ Monthly rotation documented
└─ Secure secret storage required

Secrets Storage:

Critical - Use Secure Vault:
├─ HashiCorp Vault (recommended)
├─ AWS Secrets Manager
├─ Azure Key Vault
├─ Google Cloud Secrets
└─ Store & access policies:
   ├─ Encrypt at rest
   ├─ Encrypt in transit
   ├─ Audit all accesses
   ├─ RBAC (role-based)
   ├─ Rotation tracking
   └─ Alert on unauthorized access

Configuration:

Environment-Specific:
├─ Production region only
├─ Performance settings optimized
├─ Logging: Moderate (balance vs storage)
├─ Monitoring: Comprehensive
├─ Alerting: All channels
└─ Recovery procedures: Documented

CI/CD Pipeline:

Strict Requirements:
├─ All tests: Must pass
├─ Code review: Mandatory
├─ Approval: Required (2+ reviewers)
├─ Build: Automated & verified
├─ Deploy: Gated release
├─ Smoke tests: Post-deploy (critical)
├─ Rollback: Instant available
└─ Notify: Multiple teams

Deployment Procedure:

1. Code Merge (to main branch)
   ├─ Tests pass
   ├─ Reviews approved
   └─ CI/CD starts

2. Build (automated)
   ├─ Code compiled
   ├─ Tests run
   ├─ Artifact created
   └─ Ready for deploy

3. Stage (automated)
   ├─ Deploy to staging
   ├─ Smoke tests run
   ├─ If all pass: await approval
   └─ If any fail: block deployment

4. Approve (manual gate)
   ├─ Engineering manager: approval
   ├─ On-call engineer: ready
   ├─ Time window: Business hours preferred
   └─ Cancellation: Available anytime

5. Deploy (automated)
   ├─ Deploy to production
   ├─ Gradual rollout (optional)
   ├─ Monitor deployment
   ├─ Health checks pass
   └─ Notify team

6. Monitor (continuous)
   ├─ Watch logs
   ├─ Watch metrics
   ├─ Watch errors
   ├─ Alert threshold: Low
   └─ Quick response ready

7. Rollback (if needed)
   ├─ Detected: Issue identified
   ├─ Decision: Rollback authorized
   ├─ Execute: One command
   ├─ Verify: Health checks pass
   └─ Notify: Team aware

Monitoring:

Comprehensive Observability:
├─ Application logging (ELK/Splunk)
├─ Error tracking (Sentry/Rollbar)
├─ APM (New Relic/DataDog)
├─ Uptime monitoring (Synthetic)
├─ Custom metrics
├─ Infrastructure monitoring
└─ Security monitoring

Alerting:

Critical (Immediate - Page):
├─ Application down (503/504)
├─ Authentication failures spike
├─ Database connection errors
├─ Memory/CPU critically high
├─ Deployment failed
└─ Security incident

High Priority (30 min - Slack):
├─ Error rate > 5%
├─ Response time spike
├─ Rate limit approached
├─ Token refresh failures
├─ API quota exceeded
└─ Unusual traffic pattern

Medium Priority (1 hour - Email):
├─ Minor errors
├─ Performance degradation
├─ Deprecated API usage
├─ Scheduled job delayed
└─ Configuration drift

Disaster Recovery:

Backup & Recovery:
├─ Database: Daily snapshots
├─ Configuration: Version controlled
├─ Secrets: Vault with audit logs
├─ RTO: < 1 hour
├─ RPO: < 15 minutes
└─ Tested monthly

Communication:

During Incident:
├─ Status page: Updated immediately
├─ Slack: Team notification
├─ Email: If long duration
├─ Customers: If customer-facing
└─ Escalation: Clear path

Post-Incident:
├─ Post-mortem: Scheduled
├─ Root cause: Identified
├─ Fix: Implemented
├─ Prevention: For future
└─ Lessons learned: Documented

Compliance:

Security Requirements:
├─ PCI-DSS: For payment data
├─ HIPAA: For health data
├─ SOC 2: For SaaS
├─ GDPR: For EU data
└─ Industry-specific: As applicable

Audit & Compliance:
├─ Audit logs: Retained 2+ years
├─ Access logs: Who did what when
├─ Change logs: All changes recorded
├─ Compliance checks: Quarterly
└─ External audits: Annual

Troubleshooting Deployments

Common Issues:

Problem: Authentication Fails in Production

Causes:
├─ Secret rotated, app not updated
├─ Secret expired (if revoked)
├─ Client deleted
├─ Secret wrong (typo)
└─ Wrong client for environment

Diagnosis:
├─ Check error: Invalid credentials
├─ Verify secret matches what stored
├─ Verify client still exists
├─ Check expiration date
└─ Try authentication manually

Fix:
├─ If secret wrong: Update immediately
├─ If client deleted: Recreate
├─ If expired: Generate new secret
├─ If revoked: Verify it's revoked
└─ Test after fix

Prevention:
├─ Document secret location
├─ Rotate on schedule
├─ Test auth before deploy
├─ Use vault for storage
└─ Alert before expiration

Problem: Rate Limiting in Production

Symptoms:
├─ 429 errors increasing
├─ Request latency increasing
├─ API calls slowing down
├─ Partial failures
└─ Customers reporting issues

Diagnosis:
├─ Check request volume
├─ Check request frequency
├─ Identify hot endpoints
├─ Check X-Rate-Limit-Remaining
└─ Calculate current rate

Fix (Short-term):
├─ Implement backoff
├─ Reduce request frequency
├─ Queue requests
├─ Reduce batch size
└─ Wait for window reset

Fix (Long-term):
├─ Use bulk endpoints
├─ Implement caching
├─ Use WebSocket events
├─ Optimize queries
└─ Consider enterprise limits

Prevention:
├─ Load test before production
├─ Monitor rate limit headers
├─ Alert when approaching limit
├─ Design for pagination
└─ Use bulk APIs from start

Problem: Token Refresh Failing

Symptoms:
├─ API calls return 401
├─ Refresh token errors
├─ Users getting logged out
├─ Authentication failing
└─ Errors in logs

Diagnosis:
├─ Check refresh token expired
├─ Check client exists
├─ Check secret correct
├─ Check token lifetime
└─ Try manual refresh

Fix:
├─ If expired: User must re-authenticate
├─ If client wrong: Update config
├─ If secret wrong: Update secret
├─ If lifetime: Adjust config
└─ Test refresh manually

Prevention:
├─ Refresh proactively (5 min before)
├─ Monitor refresh success rate
├─ Alert on refresh failures
├─ Document refresh logic
└─ Test token refresh

Problem: Data Loss / Sync Issues

Symptoms:
├─ Contacts not syncing
├─ Data inconsistencies
├─ Partial updates missing
├─ Database conflicts
└─ Customers reporting issues

Diagnosis:
├─ Check sync logs
├─ Verify API calls succeeded
├─ Check for conflict errors
├─ Verify data format
└─ Compare source vs target

Fix:
├─ Re-run sync
├─ Correct data format
├─ Handle conflicts
├─ Verify completeness
└─ Reconcile manually if needed

Prevention:
├─ Implement idempotency
├─ Use external IDs
├─ Log all changes
├─ Verify sync completion
├─ Regular reconciliation
├─ Alerts on failures
└─ Testing with real data

Problem: Performance Degradation

Symptoms:
├─ Slow API response times
├─ High latency (p99)
├─ User complaints
├─ Dashboard sluggish
└─ Timeout errors

Diagnosis:
├─ Check API latency
├─ Check application latency
├─ Check database latency
├─ Check network latency
└─ Identify bottleneck

Fix (Short-term):
├─ Scale up application
├─ Scale up database
├─ Clear cache
├─ Reduce requests
└─ Optimize queries

Fix (Long-term):
├─ Use caching layer
├─ Optimize database
├─ Use CDN
├─ Async processing
└─ Horizontal scaling

Prevention:
├─ Load test regularly
├─ Monitor latency
├─ Alert on degradation
├─ Capacity planning
└─ Performance budget

Key Takeaways: Chapter 8

• Pattern Diversity - Different patterns for different scenarios
• Authentication Clear - Always OAuth 2.0 (Client Credentials or Code)
• Scheduling Critical - Cron jobs for reliable automation
• Error Handling - Implement retry logic with backoff
• Data Quality - Validate and normalize before syncing
• Deployment Gates - Approval required for production
• Monitoring Essential - Know when things break
• Documentation Important - For debugging and maintenance

Interview Prep: Integration Patterns

Document Version

Chapter: 8 of 8
Last Updated: March 2026
Status: Current with OAuth 2.0 & API standards
Scope: Integration patterns, deployment strategies, troubleshooting