4. Application Cutover

Execute a smooth, low-downtime transition from your legacy infrastructure to CloudNexus. This step-by-step guide covers DNS switching, traffic routing, and rollback procedures.

ℹ️
Prerequisites

Before starting the cutover, ensure Steps 1–3 (Environment Assessment, Infrastructure Provisioning, and Data Migration) are fully completed and validated. Your CloudNexus environment should be fully provisioned and your data synchronized.

The application cutover is the critical phase where live traffic transitions from your existing infrastructure to CloudNexus. Our recommended approach uses a phased cutover strategy to minimize downtime and risk. Typical cutover windows range from 15 to 60 minutes depending on your architecture complexity.

Cutover Strategy Overview

CloudNexus supports three cutover strategies depending on your risk tolerance and SLA requirements:

Strategy Downtime Complexity Best For
Big Bang 15–30 min Low Simple apps, maintenance windows
Phased (Recommended) 0–5 min Medium Most production workloads
Blue-Green Zero High Zero-downtime requirements
Recommended Approach

For most production environments, we recommend the Phased Cutover strategy. It provides a good balance between speed, safety, and operational complexity. The following steps detail this approach.

Pre-Cutover Checklist

Complete every item on this checklist before initiating the cutover window. Any uncompleted items should be escalated to your CloudNexus Solutions Architect.

  • Infrastructure provisioned and health-checked on CloudNexus
  • Data migration completed with final delta sync verified
  • Application deployed and smoke tests passing on CloudNexus
  • DNS records identified and TTLs reduced to 300s (1 hour prior)
  • Load balancer configured and health checks passing
  • SSL/TLS certificates provisioned and validated
  • Monitoring dashboards and alerting configured
  • Rollback procedure documented and tested
  • Stakeholder notification sent to all teams
  • War room / bridge call established with key personnel

Cutover Execution Steps

The following steps should be executed sequentially during the cutover window. Each step includes validation criteria before proceeding to the next.

1
Place Origin in Maintenance Mode
⏱ ~2 min

Enable maintenance mode on your origin infrastructure to prevent new write operations. This ensures data consistency during the final sync.

Use the CloudNexus CLI to trigger maintenance mode on your load balancer:

Bash
# Enable maintenance mode on origin load balancer cnx lb set-maintenance --origin-pool legacy-prod-pool --status enabled # Verify maintenance mode is active cnx lb get-status --origin-pool legacy-prod-pool { "maintenance_mode": true, "active_connections": 0 }]

Validation: Confirm that active_connections equals 0 and the origin is returning HTTP 503 responses.

2
Execute Final Data Sync
⏱ ~5–10 min

Run the final delta synchronization to capture any data changes that occurred during the maintenance mode setup. This is the last data transfer before cutover.

Bash
# Run final delta sync cnx migrate sync-final --job-id mig-7f3a9b2c --mode delta # Verify sync completion cnx migrate verify-checksum --job-id mig-7f3a9b2c ✓ Source checksum: a3f8c2d1e4b7... Target checksum: a3f8c2d1e4b7... Status: MATCH
🚨
Critical — Checksum Mismatch

If the checksums do not match, abort the cutover and re-run the sync. Do not proceed with traffic switching until data integrity is confirmed.

Validation: Checksums must match between source and target. All replication lag metrics should read 0.

3
Update DNS Records
⏱ ~2 min

Update your DNS A/AAAA records to point to your CloudNexus global load balancer IP. Since TTLs were reduced earlier, propagation will occur within minutes.

Bash
# Update DNS to CloudNexus load balancer cnx dns update-record --domain app.yourcompany.com \ --type A --value 198.51.100.42 --ttl 300 # For IPv6 cnx dns update-record --domain app.yourcompany.com \ --type AAAA --value 2001:db8::42 --ttl 300 # Verify DNS propagation cnx dns check-propagation --domain app.yourcompany.com ✓ US-East: 198.51.100.42 (propagated) ✓ EU-West: 198.51.100.42 (propagated) ✓ AP-South: 198.51.100.42 (propagated) ✓ SA-East: 198.51.100.42 (propagated)

Validation: All global DNS locations must resolve to the CloudNexus IP address.

4
Activate CloudNexus Load Balancer
⏱ ~2 min

Switch the CloudNexus load balancer from drain mode to active, beginning to accept and route live traffic to your application.

Bash
# Activate the CloudNexus load balancer cnx lb activate --lb-id lb-cn-x9k2m --mode accepting # Enable WAF and DDoS protection cnx security enable --lb-id lb-cn-x9k2m --waf --ddos # Verify health checks cnx lb health-check --lb-id lb-cn-x9k2m ✓ us-east-1a app-node-01 healthy latency: 2ms ✓ us-east-1b app-node-02 healthy latency: 3ms ✓ us-west-2a app-node-03 healthy latency: 4ms

Validation: All backend nodes report healthy status with latency under 10ms.

5
Phased Traffic Shift
⏱ ~10–20 min

Gradually shift traffic from the origin to CloudNexus using weighted routing. This allows you to monitor application behavior under real load and catch issues early.

Bash
# Phase 1: Shift 10% traffic to CloudNexus cnx lb set-weight --lb-id lb-cn-x9k2m --cnx 10 --origin 90 # Wait 5 minutes and verify metrics cnx monitor check-errors --lb-id lb-cn-x9k2m --window 5m Error rate: 0.02% (within threshold) P99 latency: 45ms (within threshold) # Phase 2: Shift 50% traffic cnx lb set-weight --lb-id lb-cn-x9k2m --cnx 50 --origin 50 # Phase 3: Shift 100% traffic cnx lb set-weight --lb-id lb-cn-x9k2m --cnx 100 --origin 0

Validation: At each phase, verify error rates are below 0.1% and latency is within acceptable thresholds. If errors spike, hold at the current weight and investigate.

6
Disable Origin & Finalize
⏱ ~3 min

Once 100% of traffic is confirmed flowing through CloudNexus, disable the origin infrastructure and remove it from the load balancing pool.

Bash
# Remove origin from load balancer pool cnx lb remove-pool --lb-id lb-cn-x9k2m --pool legacy-prod-pool # Disable origin maintenance mode (no longer needed) cnx lb set-maintenance --origin-pool legacy-prod-pool --status disabled # Run final cutover verification cnx migrate cutover-complete --job-id mig-7f3a9b2c ✓ DNS propagation: 100% complete ✓ Load balancer: Active, accepting traffic ✓ Error rate: 0.01% ✓ Data consistency: Verified ✓ SSL certificates: Valid 🎉 Cutover completed successfully!

Validation: The cutover-complete command must return all green checkmarks. Archive the cutover log for your records.

Expected Cutover Timeline

T-0:00
War Room Opens
All stakeholders join the bridge call. Roles and responsibilities confirmed.
T+0:02
Maintenance Mode Activated
Origin infrastructure placed in maintenance mode. Write operations blocked.
T+0:12
Final Data Sync Complete
Delta sync finished, checksums verified, data integrity confirmed.
T+0:15
DNS Updated
A/AAAA records updated. Propagation begins immediately (TTL = 300s).
T+0:18
Traffic Shift Begins
10% → 50% → 100% traffic shift with monitoring at each phase.
T+0:35
Full Traffic on CloudNexus
100% traffic confirmed. Origin removed from pool.
T+0:40
Cutover Complete
Final verification passed. War room stands down. Monitoring continues.

Rollback Procedures

⚠️
When to Rollback

Initiate a rollback if any of the following occur: error rate exceeds 1%, P99 latency exceeds 2× baseline, data corruption is detected, or critical features are non-functional on CloudNexus.

If a rollback is required, execute the following steps in order. The goal is to restore traffic to the origin as quickly as possible.

Bash
# ROLLBACK: Shift all traffic back to origin cnx lb set-weight --lb-id lb-cn-x9k2m --cnx 0 --origin 100 # Revert DNS records to origin IPs cnx dns rollback --domain app.yourcompany.com --job-id mig-7f3a9b2c # Remove maintenance mode from origin cnx lb set-maintenance --origin-pool legacy-prod-pool --status disabled # Verify origin is accepting traffic cnx lb health-check --origin-pool legacy-prod-pool ✓ origin-node-01 healthy latency: 12ms ✓ origin-node-02 healthy latency: 14ms ⚠️ Rollback complete. Traffic restored to origin.

Important: After a rollback, the CloudNexus environment remains provisioned and ready. Resolve the identified issues and reschedule the cutover. Your Solutions Architect will conduct a post-mortem to determine root cause.

Post-Cutover Monitoring

After a successful cutover, maintain elevated monitoring for at least 2 hours. Key metrics to watch:

  • HTTP Error Rates: Must remain below 0.1%. Watch for 5xx errors specifically.
  • Latency Percentiles: P95 and P99 should be within 10% of pre-cutover baselines.
  • Database Connection Pool: Monitor for connection spikes or exhaustions.
  • Cache Hit Rates: Verify that CDN and application caches are warming correctly.
  • Application Logs: Watch for new error patterns or warnings in CloudWatch / Datadog.
  • Business Metrics: Conversion rates, API throughput, and user session data.

CloudNexus provides a Cutover Dashboard accessible from your console that aggregates all these metrics into a single view. Your Solutions Architect can add custom dashboards for application-specific KPIs.

🎉
Cutover Successful!

If all monitoring checks pass after 2 hours, your cutover is considered successful. Proceed to Step 5: Post-Migration Validation for comprehensive testing and optimization.

Best Practices

Schedule During Low Traffic

Plan your cutover during your lowest traffic period, typically between 2:00 AM – 5:00 AM in your primary region. Use historical traffic data from the CloudNexus migration assessment to identify the optimal window.

Reduce TTLs Early

Reduce DNS TTLs to 300 seconds (5 minutes) at least 1 hour before the cutover window. This ensures rapid propagation when records are updated. Do not go below 300s as some recursive DNS resolvers may ignore lower values.

Test the Rollback

Perform a dry-run of the rollback procedure during a staging migration. Verify that traffic can be restored to the origin in under 5 minutes.

Communicate Proactively

Notify all stakeholders (engineering, support, marketing, executive team) at least 48 hours before the cutover. Include the expected downtime window, rollback criteria, and contact information for the war room.

Keep the Origin Warm

Do not decommission the origin infrastructure immediately after cutover. Keep it running in a reduced state for at least 7 days as a rollback safety net.